You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 

1163 lines
53 KiB

{
"cells": [
{
"cell_type": "markdown",
"id": "a98030af-fcd1-4d63-a36e-38ba053498fa",
"metadata": {},
"source": [
"# A full business solution\n",
"\n",
"Create a product that builds a Brochure for a company to be used for prospective clients, investors and potential recruits.\n",
"\n",
"We will be provided a company name and their primary website."
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "d5b08506-dc8b-4443-9201-5f1848161363",
"metadata": {},
"outputs": [],
"source": [
"# imports\n",
"\n",
"import os\n",
"import requests\n",
"import json\n",
"from typing import List\n",
"from dotenv import load_dotenv\n",
"from bs4 import BeautifulSoup\n",
"from IPython.display import Markdown, display, update_display\n",
"from openai import OpenAI"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "fc5d8880-f2ee-4c06-af16-ecbc0262af61",
"metadata": {},
"outputs": [],
"source": [
"# Initialize and constants\n",
"\n",
"load_dotenv()\n",
"os.environ['OPENAI_API_KEY'] = os.getenv('OPENAI_API_KEY', 'your-key-if-not-using-env')\n",
"MODEL = 'gpt-4o-mini'\n",
"openai = OpenAI()"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "106dd65e-90af-4ca8-86b6-23a41840645b",
"metadata": {},
"outputs": [],
"source": [
"# A class to represent a Webpage\n",
"\n",
"class Website:\n",
" url: str\n",
" title: str\n",
" body: str\n",
" links: List[str]\n",
"\n",
" def __init__(self, url):\n",
" self.url = url\n",
" response = requests.get(url)\n",
" self.body = response.content\n",
" soup = BeautifulSoup(self.body, 'html.parser')\n",
" self.title = soup.title.string if soup.title else \"No title found\"\n",
" if soup.body:\n",
" for irrelevant in soup.body([\"script\", \"style\", \"img\", \"input\"]):\n",
" irrelevant.decompose()\n",
" self.text = soup.body.get_text(separator=\"\\n\", strip=True)\n",
" else:\n",
" self.text = \"\"\n",
" links = [link.get('href') for link in soup.find_all('a')]\n",
" self.links = [link for link in links if link]\n",
"\n",
" def get_contents(self):\n",
" return f\"Webpage Title:\\n{self.title}\\nWebpage Contents:\\n{self.text}\\n\\n\""
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "e30d8128-933b-44cc-81c8-ab4c9d86589a",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Webpage Title:\n",
"Home - Edward Donner\n",
"Webpage Contents:\n",
"Home\n",
"Outsmart\n",
"An arena that pits LLMs against each other in a battle of diplomacy and deviousness\n",
"About\n",
"Posts\n",
"Well, hi there.\n",
"I’m Ed. I like writing code and experimenting with LLMs, and hopefully you’re here because you do too. I also enjoy DJing (but I’m badly out of practice), amateur electronic music production (\n",
"very\n",
"amateur) and losing myself in\n",
"Hacker News\n",
", nodding my head sagely to things I only half understand.\n",
"I’m the co-founder and CTO of\n",
"Nebula.io\n",
". We’re applying AI to a field where it can make a massive, positive impact: helping people discover their potential and pursue their reason for being. Recruiters use our product today to source, understand, engage and manage talent. I’m previously the founder and CEO of AI startup untapt,\n",
"acquired in 2021\n",
".\n",
"We work with groundbreaking, proprietary LLMs verticalized for talent, we’ve\n",
"patented\n",
"our matching model, and our award-winning platform has happy customers and tons of press coverage.\n",
"Connect\n",
"with me for more!\n",
"August 6, 2024\n",
"Outsmart LLM Arena – a battle of diplomacy and deviousness\n",
"June 26, 2024\n",
"Choosing the Right LLM: Toolkit and Resources\n",
"February 7, 2024\n",
"Fine-tuning an LLM on your texts: a simulation of you\n",
"January 31, 2024\n",
"Fine-tuning an LLM on your texts: part 4 – QLoRA\n",
"Navigation\n",
"Home\n",
"Outsmart\n",
"An arena that pits LLMs against each other in a battle of diplomacy and deviousness\n",
"About\n",
"Posts\n",
"Get in touch\n",
"ed [at] edwarddonner [dot] com\n",
"www.edwarddonner.com\n",
"Follow me\n",
"LinkedIn\n",
"Twitter\n",
"Facebook\n",
"Subscribe to newsletter\n",
"Type your email…\n",
"Subscribe\n",
"\n",
"\n"
]
}
],
"source": [
"ed = Website(\"https://edwarddonner.com\")\n",
"print(ed.get_contents())"
]
},
{
"cell_type": "markdown",
"id": "1771af9c-717a-4fca-bbbe-8a95893312c3",
"metadata": {},
"source": [
"## First step: Have GPT-4o-mini figure out which links are relevant\n",
"\n",
"### Use a call to gpt-4o-mini to read the links on a webpage, and respond in structured JSON. \n",
"It should decide which links are relevant, and replace relative links such as \"/about\" with \"https://company.com/about\". \n",
"We will use \"one shot prompting\" in which we provide an example of how it should respond in the prompt."
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "6957b079-0d96-45f7-a26a-3487510e9b35",
"metadata": {},
"outputs": [],
"source": [
"link_system_prompt = \"You are provided with a list of links found on a webpage. \\\n",
"You are able to decide which of the links would be most relevant to include in a brochure about the company, \\\n",
"such as links to an About page, or a Company page, or Careers/Jobs pages.\\n\"\n",
"link_system_prompt += \"You should respond in JSON as in this example:\"\n",
"link_system_prompt += \"\"\"\n",
"{\n",
" \"links\": [\n",
" {\"type\": \"about page\", \"url\": \"https://full.url/goes/here/about\"},\n",
" {\"type\": \"careers page\": \"url\": \"https://another.full.url/careers\"}\n",
" ]\n",
"}\n",
"\"\"\""
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "8e1f601b-2eaf-499d-b6b8-c99050c9d6b3",
"metadata": {},
"outputs": [],
"source": [
"def get_links_user_prompt(website):\n",
" user_prompt = f\"Here is the list of links on the website of {website.url} - \"\n",
" user_prompt += \"please decide which of these are relevant web links for a brochure about the company, respond with the full https URL in JSON format. \\\n",
"Do not include Terms of Service, Privacy, email links.\\n\"\n",
" user_prompt += \"Links (some might be relative links):\\n\"\n",
" user_prompt += \"\\n\".join(website.links)\n",
" return user_prompt"
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "6bcbfa78-6395-4685-b92c-22d592050fd7",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Here is the list of links on the website of https://edwarddonner.com - please decide which of these are relevant web links for a brochure about the company, respond with the full https URL in JSON format. Do not include Terms of Service, Privacy, email links.\n",
"Links (some might be relative links):\n",
"https://edwarddonner.com/\n",
"https://edwarddonner.com/outsmart/\n",
"https://edwarddonner.com/about-me-and-about-nebula/\n",
"https://edwarddonner.com/posts/\n",
"https://edwarddonner.com/\n",
"https://news.ycombinator.com\n",
"https://nebula.io/?utm_source=ed&utm_medium=referral\n",
"https://www.prnewswire.com/news-releases/wynden-stark-group-acquires-nyc-venture-backed-tech-startup-untapt-301269512.html\n",
"https://patents.google.com/patent/US20210049536A1/\n",
"https://www.linkedin.com/in/eddonner/\n",
"https://edwarddonner.com/2024/08/06/outsmart/\n",
"https://edwarddonner.com/2024/08/06/outsmart/\n",
"https://edwarddonner.com/2024/06/26/choosing-the-right-llm-resources/\n",
"https://edwarddonner.com/2024/06/26/choosing-the-right-llm-resources/\n",
"https://edwarddonner.com/2024/02/07/fine-tune-llm-on-texts-a-simulation-of-you/\n",
"https://edwarddonner.com/2024/02/07/fine-tune-llm-on-texts-a-simulation-of-you/\n",
"https://edwarddonner.com/2024/01/31/fine-tuning-an-llm-on-your-text-messages-using-qlora/\n",
"https://edwarddonner.com/2024/01/31/fine-tuning-an-llm-on-your-text-messages-using-qlora/\n",
"https://edwarddonner.com/\n",
"https://edwarddonner.com/outsmart/\n",
"https://edwarddonner.com/about-me-and-about-nebula/\n",
"https://edwarddonner.com/posts/\n",
"mailto:hello@mygroovydomain.com\n",
"https://www.linkedin.com/in/eddonner/\n",
"https://twitter.com/edwarddonner\n",
"https://www.facebook.com/edward.donner.52\n"
]
}
],
"source": [
"print(get_links_user_prompt(ed))"
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "a29aca19-ca13-471c-a4b4-5abbfa813f69",
"metadata": {},
"outputs": [],
"source": [
"def get_links(url):\n",
" website = Website(url)\n",
" completion = openai.chat.completions.create(\n",
" model=MODEL,\n",
" messages=[\n",
" {\"role\": \"system\", \"content\": link_system_prompt},\n",
" {\"role\": \"user\", \"content\": get_links_user_prompt(website)}\n",
" ],\n",
" response_format={\"type\": \"json_object\"}\n",
" )\n",
" result = completion.choices[0].message.content\n",
" return json.loads(result)"
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "d3d583e2-dcc4-40cc-9b28-1e8dbf402924",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"{'links': [{'type': 'about page', 'url': 'https://anthropic.com/company'},\n",
" {'type': 'careers page', 'url': 'https://anthropic.com/careers'},\n",
" {'type': 'team page', 'url': 'https://anthropic.com/team'},\n",
" {'type': 'research page', 'url': 'https://anthropic.com/research'},\n",
" {'type': 'news page', 'url': 'https://anthropic.com/news'}]}"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"get_links(\"https://anthropic.com\")"
]
},
{
"cell_type": "markdown",
"id": "0d74128e-dfb6-47ec-9549-288b621c838c",
"metadata": {},
"source": [
"## Second step: make the brochure!\n",
"\n",
"Assemble all the details into another prompt to GPT4-o"
]
},
{
"cell_type": "code",
"execution_count": 12,
"id": "85a5b6e2-e7ef-44a9-bc7f-59ede71037b5",
"metadata": {},
"outputs": [],
"source": [
"def get_all_details(url):\n",
" result = \"Landing page:\\n\"\n",
" result += Website(url).get_contents()\n",
" links = get_links(url)\n",
" print(\"Found links:\", links)\n",
" for link in links[\"links\"]:\n",
" result += f\"\\n\\n{link['type']}\\n\"\n",
" result += Website(link[\"url\"]).get_contents()\n",
" return result"
]
},
{
"cell_type": "code",
"execution_count": 13,
"id": "5099bd14-076d-4745-baf3-dac08d8e5ab2",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found links: {'links': [{'type': 'about page', 'url': 'https://anthropic.com/company'}, {'type': 'careers page', 'url': 'https://anthropic.com/careers'}, {'type': 'team page', 'url': 'https://anthropic.com/team'}, {'type': 'research page', 'url': 'https://anthropic.com/research'}]}\n",
"Landing page:\n",
"Webpage Title:\n",
"Home \\ Anthropic\n",
"Webpage Contents:\n",
"Claude\n",
"Overview\n",
"Team\n",
"API\n",
"Pricing\n",
"Research\n",
"Company\n",
"Careers\n",
"News\n",
"AI\n",
"research\n",
"and\n",
"products\n",
"that put safety at the frontier\n",
"New\n",
"Meet Claude 3.5 Sonnet\n",
"Claude 3.5 Sonnet, our most intelligent AI model, is now available.\n",
"Talk to Claude\n",
"API\n",
"Build with Claude\n",
"Start using Claude to drive efficiency and create new revenue streams.\n",
"Get started now\n",
"Our Work\n",
"See All\n",
"Announcements\n",
"Claude 3.5 Sonnet\n",
"Jun 21, 2024\n",
"Alignment\n",
"·\n",
"Research\n",
"Constitutional AI: Harmlessness from AI Feedback\n",
"Dec 15, 2022\n",
"Announcements\n",
"Core Views on AI Safety: When, Why, What, and How\n",
"Mar 8, 2023\n",
"Work with Anthropic\n",
"Anthropic is an AI safety and research company based in San Francisco. Our interdisciplinary team has experience across ML, physics, policy, and product. Together, we generate research and create reliable, beneficial AI systems.\n",
"See open roles\n",
"Claude\n",
"API\n",
"Team\n",
"Pricing\n",
"Research\n",
"Company\n",
"Customers\n",
"News\n",
"Careers\n",
"Press Inquiries\n",
"Support\n",
"Status\n",
"Twitter\n",
"LinkedIn\n",
"Availability\n",
"Terms of Service – Consumer\n",
"Terms of Service – Commercial\n",
"Privacy Policy\n",
"Usage Policy\n",
"Responsible Disclosure Policy\n",
"Compliance\n",
"Privacy Choices\n",
"© 2024 Anthropic PBC\n",
"\n",
"\n",
"\n",
"about page\n",
"Webpage Title:\n",
"Company \\ Anthropic\n",
"Webpage Contents:\n",
"Claude\n",
"Overview\n",
"Team\n",
"API\n",
"Pricing\n",
"Research\n",
"Company\n",
"Careers\n",
"News\n",
"Making AI systems\n",
"you can rely on\n",
"Anthropic is an AI safety and research company. We build reliable, interpretable, and steerable AI systems.\n",
"Join us\n",
"Our Purpose\n",
"We believe AI will have a vast impact on the world. Anthropic is dedicated to building systems that people can rely on and generating research about the opportunities and risks of AI.\n",
"We Build Safer Systems\n",
"We aim to build frontier AI systems that are reliable, interpretable, and steerable. We conduct frontier research, develop and apply a variety of safety techniques, and deploy the resulting systems via a set of partnerships and products.\n",
"Safety Is a Science\n",
"We treat AI safety as a systematic science, conducting research, applying it to our products, feeding those insights back into our research, and regularly sharing what we learn with the world along the way.\n",
"Interdisciplinary\n",
"Anthropic is a collaborative team of researchers, engineers, policy experts, business leaders and operators, who bring our experience from many different domains to our work.\n",
"AI Companies are One Piece of a Big Puzzle\n",
"AI has the potential to fundamentally change how the world works. We view ourselves as just one piece of this evolving puzzle. We collaborate with civil society, government, academia, nonprofits and industry to promote safety industry-wide.\n",
"The Team\n",
"We’re a team of researchers, engineers, policy experts and operational leaders, with experience spanning a variety of disciplines, all working together to build reliable and understandable AI systems.\n",
"Research\n",
"We conduct frontier AI research across a variety of modalities, and explore novel and emerging safety research areas from interpretability to RL from human feedback to policy and societal impacts analysis.\n",
"Policy\n",
"We think about the impacts of our work and strive to communicate what we’re seeing at the frontier to policymakers and civil society in the US and abroad to help promote safe and reliable AI.\n",
"Product\n",
"We translate our research into tangible, practical tools like Claude that benefit businesses, nonprofits and civil society groups and their clients and people around the globe.\n",
"Operations\n",
"Our people, finance, legal, and recruiting teams are the human engines that make Anthropic go. We’ve had previous careers at NASA, startups, and the armed forces and our diverse experiences help make Anthropic a great place to work (and we love plants!).\n",
"Our Values\n",
"01\n",
"Here for the mission\n",
"Anthropic exists for our mission: to ensure transformative AI helps people and society flourish. Progress this decade may be rapid, and we expect increasingly capable systems to pose novel challenges. We pursue our mission by building frontier systems, studying their behaviors, working to responsibly deploy them, and regularly sharing our safety insights. We collaborate with other projects and stakeholders seeking a similar outcome.\n",
"02\n",
"Unusually high trust\n",
"Our company is an unusually high trust environment: we assume good faith, disagree kindly, and prioritize honesty. We expect emotional maturity and intellectual openness. At its best, our trust enables us to make better decisions as an organization than any one of us could as individuals.\n",
"03\n",
"One big team\n",
"Collaboration is central to our work, culture, and value proposition. While we have many teams at Anthropic, we feel the broader sense in which we are all on the same team working together towards the mission. Leadership sets the strategy, with broad input from everyone, and trusts each piece of the organization to pursue these goals in their unique style. Individuals commonly contribute to work across many different areas.\n",
"04\n",
"Do the simple thing that works\n",
"We celebrate trying the simple thing before the clever, novel thing. We embrace pragmatism - sensible, practical approaches that acknowledge tradeoffs. We love empiricism - finding out what actually works by trying it - and apply this to our research, our engineering and our collaboration. We aim to be open about what we understand and what we don’t.\n",
"Governance\n",
"Anthropic is a Public Benefit Corporation, whose purpose is the responsible development and maintenance of advanced AI for the long-term benefit of humanity. Our Board of Directors is elected by stockholders and our Long-Term Benefit Trust, as explained\n",
"here.\n",
"Current members of the Board and the Long-Term Benefit Trust (LTBT) are listed below.\n",
"Anthropic Board of Directors\n",
"Dario Amodei, Daniela Amodei, Yasmin Razavi, and Jay Kreps.\n",
"LTBT Trustees\n",
"Neil Buddy Shah, Kanika Bahl, and Zach Robinson.\n",
"Company News\n",
"See All\n",
"Announcements\n",
"Artifacts are now generally available\n",
"Aug 27, 2024\n",
"Announcements\n",
"Expanding our model safety bug bounty program\n",
"Aug 8, 2024\n",
"Announcements\n",
"Claude is now available in Brazil\n",
"Aug 1, 2024\n",
"Want to help us build the future of safe AI?\n",
"Join us\n",
"Claude\n",
"API\n",
"Team\n",
"Pricing\n",
"Research\n",
"Company\n",
"Customers\n",
"News\n",
"Careers\n",
"Press Inquiries\n",
"Support\n",
"Status\n",
"Twitter\n",
"LinkedIn\n",
"Availability\n",
"Terms of Service – Consumer\n",
"Terms of Service – Commercial\n",
"Privacy Policy\n",
"Usage Policy\n",
"Responsible Disclosure Policy\n",
"Compliance\n",
"Privacy Choices\n",
"© 2024 Anthropic PBC\n",
"\n",
"\n",
"\n",
"careers page\n",
"Webpage Title:\n",
"Careers \\ Anthropic\n",
"Webpage Contents:\n",
"Claude\n",
"Overview\n",
"Team\n",
"API\n",
"Pricing\n",
"Research\n",
"Company\n",
"Careers\n",
"News\n",
"Join the team\n",
"making AI safe\n",
"We’re a public benefit corporation headquartered in San Francisco. Our team’s experience spans a variety of backgrounds and disciplines, from physics and machine learning to public policy and business. We work as a cohesive team that collectively forecasts the impact and tractability of research ideas in advancing our mission.\n",
"See open roles\n",
"What We Offer\n",
"Health & Wellness\n",
"We offer a range of benefits to best support your and your family's wellbeing.\n",
"Comprehensive health, dental, and vision insurance for you and your dependents\n",
"Inclusive fertility benefits via Carrot Fertility\n",
"Generous subsidy for OneMedical\n",
"22 weeks of paid parental leave\n",
"Unlimited PTO – most staff take between 4-6 weeks each year, sometimes more\n",
"Compensation & Support\n",
"We offer competitive compensation with significant amounts of equity. Your equity can be multiplied if you choose to donate a portion of it to charity.\n",
"Competitive salary and equity packages\n",
"Optional equity donation matching at a 1:1 ratio, up to 25% of your equity grant\n",
"401(k) plan with 4% matching\n",
"Additional Benefits\n",
"We’re continually upgrading our benefits program so we can meet the needs of our entire team.\n",
"$500/month flexible wellness stipend\n",
"Commuter coverage\n",
"Annual education stipend\n",
"A home office improvement stipend when you first join\n",
"Relocation support for those moving to the Bay Area\n",
"Daily lunches in the office\n",
"How We Hire\n",
"The interview process at Anthropic varies based on role and candidate, but our standard process looks like this:\n",
"Step 1\n",
"Resume\n",
"Submit your resume via our website.\n",
"Step 2\n",
"Exploratory chat\n",
"You’ll have a chat with one of our staff to discuss your career interests and relevant experience, and learn more about Anthropic.\n",
"Step 3\n",
"Skills Assessment\n",
"For technical roles, you’ll have a one-hour technical screening interview.\n",
"For operations or policy roles, you’ll get a take-home assignment. These typically involve writing responses to several role-relevant questions; they may occasionally require some outside research. Assignments usually take between 2-5 hours, depending on the role.\n",
"We include this to minimize bias and make well-informed hiring decisions. We think seeing a candidate’s work helps us assess how they might actually perform on the job; similarly, the assignment gives candidates a better idea of what their work at Anthropic might entail. If a candidate likes working through their take-home, that is one indicator that they would enjoy taking on the role, and vice versa.\n",
"We recognize that completing work assignments requires time and effort, and that they are not perfectly reflective of the role’s work. Nonetheless, we think that work tests are a useful complement to interviews and reference checks.\n",
"Step 4\n",
"Team Screen\n",
"You'll have a conversation with either the Hiring Manager or a member of your potential team.\n",
"Step 5\n",
"Interview Panel\n",
"For technical roles, you’ll have 3-4 more one-hour technical interviews, plus a culture interview.\n",
"For operations or policy roles, you’ll have 3-5 hours of interviews, including a culture interview.\n",
"Step 6\n",
"Final Checks\n",
"We’ll ask for some references, and have you chat with our leadership.\n",
"Step 7\n",
"Offer\n",
"We’ll make you an offer!\n",
"Technical Interviews\n",
"Technical interviews at Anthropic are broadly categorized into ‘engineering’ or ‘research’ interviews, and each candidate is given a mix tailored to their skillset.\n",
"Engineering interviews are usually carried out in a shared Python coding environment, like Google Colab. Frontend engineering interviews are in JavaScript. They have the form:\n",
"Here’s a description of a component from our stack. Could you re-implement a toy version of it for me in one hour?\n",
"These components are ‘chunkier’ than the more common LeetCode problems, and are intended to mimic the day-to-day of engineering at Anthropic.\n",
"We are particularly interested in your thought process and how you attack the problem. You’ll be allowed to look things up with Google, but it’s still important to be familiar with Python syntax and the standard library. We primarily code in Python, and a common reason candidates fail interviews is that they're not fully comfortable in Python.\n",
"Only one of our engineering interviews touches on machine learning topics, and you can ask to pass on that one if you wish. You do not need to learn anything about machine learning before interviewing as an engineer at Anthropic.\n",
"Research interviews are broader in form. They’ll include some engineering interviews, and some discussions about the kinds of systems we study.\n",
"Both the research and engineering interview process also include softer questions about your experience and motivations, and time to ask us about Anthropic.\n",
"Other Things\n",
"Engineers here do lots of research, and researchers do lots of engineering\n",
"While there’s historically been a division between engineering and research in machine learning, we think that boundary has dissolved with the advent of large models. The distribution of candidates we interview is strongly bimodal in both engineering and research experience however, and we have necessarily tailored our interview structure to that.\n",
"If you’ve an engineering background, please apply as an engineer. You’ll perform much better in the interviews, and if you join you’ll have as much input to Anthropic’s direction and interests as anyone else.\n",
"As evidence towards this: all of our papers have engineers as authors, and often as first author. Research and engineering hires all share a single title - ‘Member of Technical Staff’.\n",
"We value direct evidence of ability\n",
"If you’ve done interesting independent research, written an insightful blog post, or made substantial contributions to open-source software, put that at the top of your resume!\n",
"Feedback\n",
"We do not provide feedback on resumes or interviews.\n",
"Visas\n",
"Anthropic sponsors visas! We aren't able to sponsor them for every role and every candidate; operations roles are especially difficult to support. But if we make you an offer, we will make every effort to get you into the United States, and we retain an immigration lawyer to help with this.\n",
"Green cards\n",
"Once you’re eligible, we’re also keen to sponsor green cards!\n",
"We do not require PhDs, degrees, or previous ML experience\n",
"About half of Anthropic technical staff have a PhD of some sort; about half had prior experience in ML. We have several brilliant colleagues who never went to college.\n",
"Remote interviewing\n",
"All our interviews are conducted over Google Meet. We prefer PST office hours, but we can be flexible if that’s difficult for you.\n",
"Re-applying\n",
"Similarly, if interviews don’t work out this time, you’re welcome to re-apply after 12 months, and earlier if something materially changes about your experience or skills.\n",
"Remote work\n",
"Anthropic staff all come to the office regularly. Most staff live in the Bay Area, though a few live further away and come in for one week a month. We also understand that moving can take time, so as a transitional phase some folks start while fully remote.\n",
"Offer timing\n",
"If we make an offer, we’re happy to give you time to think about it and finish up any other interview processes you’re going through.\n",
"Internships\n",
"We do not offer internships.\n",
"Candidate Privacy Policy\n",
"US Candidate Privacy Policy\n",
"UK Employee and Candidate Privacy Policy\n",
"Claude\n",
"API\n",
"Team\n",
"Pricing\n",
"Research\n",
"Company\n",
"Customers\n",
"News\n",
"Careers\n",
"Press Inquiries\n",
"Support\n",
"Status\n",
"Twitter\n",
"LinkedIn\n",
"Availability\n",
"Terms of Service – Consumer\n",
"Terms of Service – Commercial\n",
"Privacy Policy\n",
"Usage Policy\n",
"Responsible Disclosure Policy\n",
"Compliance\n",
"Privacy Choices\n",
"© 2024 Anthropic PBC\n",
"\n",
"\n",
"\n",
"team page\n",
"Webpage Title:\n",
"Team up with Claude \\ Anthropic\n",
"Webpage Contents:\n",
"Claude\n",
"Overview\n",
"Team\n",
"API\n",
"Pricing\n",
"Research\n",
"Company\n",
"Careers\n",
"News\n",
"Try Claude\n",
"Team up with Claude\n",
"Shorten the path from idea to impact with an AI assistant that taps into your team’s shared expertise.\n",
"Get started\n",
"Request demo\n",
"Easy collaboration for better outcomes\n",
"Claude doesn’t just speed up daily tasks like writing emails or docs. It’s a virtual teammate that moves work forward using your team’s knowledge.\n",
"Create with Claude\n",
"Claude can be a sounding board for your ideas, help you generate new ones, and pull insights from data in a snap.\n",
"Prime the canvas\n",
"Use Projects to ground Claude in specific knowledge that helps you produce higher-quality work with less effort.\n",
"Spark inspiration\n",
"Share your best chats with Claude across the team to spark creativity and improve your project deliverables.\n",
"Transform how you work\n",
"Claude makes work more productive—whether you need a partner for deep work, a creative collaborator, or an assistant for daily tasks.\n",
"Create with Claude\n",
"Draft and iterate on documents, code and, websites, and images alongside your chat with Artifacts.\n",
"Write and debug code\n",
"Create marketing campaigns\n",
"Draft job descriptions\n",
"Build interactive visualizations\n",
"Transform how your team works\n",
"Claude can serve as your go-to expert, empowering each team member with shared knowledge from all across the organization.\n",
"Prime the canvas\n",
"Create Projects and add knowledge so each person on the team can deliver expert-level results.\n",
"Find and summarize information faster\n",
"Use Claude as your subject-matter expert\n",
"Expand how each teammate can contribute\n",
"Spark inspiration\n",
"Share your best chats with everyone on the Project to spark better ideas, iterate on Artifacts, and move work forward.\n",
"Brainstorm on new product ideas\n",
"Discuss insights from user interviews\n",
"Collaborate on hard research questions\n",
"Every team can work with Claude\n",
"Engineering\n",
"Generate code snippets in seconds\n",
"Create clear, comprehensive docs with no effort\n",
"Get help debugging even the most complex issues\n",
"Turn product feedback into roadmap items faster\n",
"Support\n",
"Resolve customer issues in record time\n",
"Craft personalized responses effortlessly\n",
"Build a dynamic, user-friendly knowledge base\n",
"Generate insightful metrics reports instantly\n",
"Marketing\n",
"Create engaging content tailored to your audience\n",
"Segment customers with pinpoint accuracy\n",
"Analyze competitors with unparalleled depth\n",
"Optimize campaigns for maximum ROI\n",
"Sales\n",
"Customize pitches for any customer segment\n",
"Uncover hidden sales trends effortlessly\n",
"Draft compelling follow-up emails in seconds\n",
"Get comprehensive competitor insights on demand\n",
"By leveraging content from our help center in Projects, we were able to generate comprehensive standard operating procedures for our core workflows in just a few hours—a task that previously took our team weeks to complete.\n",
"Bradley Silicani\n",
"COO, Anrok\n",
"Claude Team is transforming our way of working at North Highland. Claude is a truly exceptional writer that has helped our team complete content creation and analysis tasks up to 5x faster than before—turning what was once two weeks of writing and research into minutes of work.\n",
"Luka Anic\n",
"Senior Director, Technical AI Program and Product Manager, North Highland\n",
"Generating content, completing creative tasks, and creating summarized reports is much easier than before. There are many other areas of our business—like engineering, legal, risk and compliance—where we're excited to see what Claude can do.\n",
"Olga Pirog\n",
"Head of AI Transformation, IG Group\n",
"Join the teams transforming with Claude\n",
"See Pricing\n",
"Claude\n",
"API\n",
"Team\n",
"Pricing\n",
"Research\n",
"Company\n",
"Customers\n",
"News\n",
"Careers\n",
"Press Inquiries\n",
"Support\n",
"Status\n",
"Twitter\n",
"LinkedIn\n",
"Availability\n",
"Terms of Service – Consumer\n",
"Terms of Service – Commercial\n",
"Privacy Policy\n",
"Usage Policy\n",
"Responsible Disclosure Policy\n",
"Compliance\n",
"Privacy Choices\n",
"© 2024 Anthropic PBC\n",
"\n",
"\n",
"\n",
"research page\n",
"Webpage Title:\n",
"Research \\ Anthropic\n",
"Webpage Contents:\n",
"Claude\n",
"Overview\n",
"Team\n",
"API\n",
"Pricing\n",
"Research\n",
"Company\n",
"Careers\n",
"News\n",
"Researching\n",
"at the frontier\n",
"At Anthropic, we develop large-scale AI systems, and our research teams help us to create safer, steerable, and more reliable models.\n",
"See open roles\n",
"Claude\n",
"API\n",
"Team\n",
"Pricing\n",
"Research\n",
"Company\n",
"Customers\n",
"News\n",
"Careers\n",
"Press Inquiries\n",
"Support\n",
"Status\n",
"Twitter\n",
"LinkedIn\n",
"Availability\n",
"Terms of Service – Consumer\n",
"Terms of Service – Commercial\n",
"Privacy Policy\n",
"Usage Policy\n",
"Responsible Disclosure Policy\n",
"Compliance\n",
"Privacy Choices\n",
"© 2024 Anthropic PBC\n",
"\n",
"\n"
]
}
],
"source": [
"print(get_all_details(\"https://anthropic.com\"))"
]
},
{
"cell_type": "code",
"execution_count": 22,
"id": "9b863a55-f86c-4e3f-8a79-94e24c1a8cf2",
"metadata": {},
"outputs": [],
"source": [
"system_prompt = \"You are an assistant that analyzes the contents of several relevant pages from a company website \\\n",
"and creates a short humorous, entertaining, jokey brochure about the company for prospective customers, investors and recruits. Respond in markdown.\\\n",
"Include details of company culture, customers and careers/jobs if you have the information.\""
]
},
{
"cell_type": "code",
"execution_count": 15,
"id": "6ab83d92-d36b-4ce0-8bcc-5bb4c2f8ff23",
"metadata": {},
"outputs": [],
"source": [
"def get_brochure_user_prompt(company_name, url):\n",
" user_prompt = f\"You are looking at a company called: {company_name}\\n\"\n",
" user_prompt += f\"Here are the contents of its landing page and other relevant pages; use this information to build a short brochure of the company in markdown.\\n\"\n",
" user_prompt += get_all_details(url)\n",
" user_prompt = user_prompt[:20_000] # Truncate if more than 20,000 characters\n",
" return user_prompt"
]
},
{
"cell_type": "code",
"execution_count": 16,
"id": "e44de579-4a1a-4e6a-a510-20ea3e4b8d46",
"metadata": {},
"outputs": [],
"source": [
"def create_brochure(company_name, url):\n",
" response = openai.chat.completions.create(\n",
" model=MODEL,\n",
" messages=[\n",
" {\"role\": \"system\", \"content\": system_prompt},\n",
" {\"role\": \"user\", \"content\": get_brochure_user_prompt(company_name, url)}\n",
" ],\n",
" )\n",
" result = response.choices[0].message.content\n",
" display(Markdown(result))"
]
},
{
"cell_type": "code",
"execution_count": 18,
"id": "e093444a-9407-42ae-924a-145730591a39",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found links: {'links': [{'type': 'about page', 'url': 'https://anthropic.com/company'}, {'type': 'careers page', 'url': 'https://anthropic.com/careers'}, {'type': 'team page', 'url': 'https://anthropic.com/team'}]}\n"
]
},
{
"data": {
"text/markdown": [
"# Anthropic Brochure\n",
"\n",
"---\n",
"\n",
"## Welcome to Anthropic\n",
"\n",
"At **Anthropic**, we are at the forefront of AI safety and research, dedicated to creating reliable, interpretable, and steerable AI systems that prioritize safety. Based in **San Francisco**, our interdisciplinary team—comprising researchers, engineers, policy experts, and business leaders—strives to ensure transformative AI benefits people and society as a whole.\n",
"\n",
"---\n",
"\n",
"## Our Mission\n",
"\n",
"### **Building Safer AI**\n",
"- We believe in the vast potential of AI and are committed to building systems that can be reliably used in real-world applications. \n",
"- Our research explores key areas such as interpretability, reinforcement learning from human feedback, and the societal impacts of AI.\n",
"\n",
"### **AI Safety as a Science**\n",
"- Safety isn't just a goal; it's a systematic science. Our approach integrates rigorous research, product application, and continuous feedback to create safer AI models.\n",
"\n",
"### **Collaborative Ecosystem**\n",
"- We are one piece of the larger AI landscape and actively collaborate with civil society, government, academia, and industry to promote wide-ranging safety measures in AI.\n",
"\n",
"---\n",
"\n",
"## Meet Claude\n",
"\n",
"Introducing **Claude**, our cutting-edge AI model currently at version **3.5 Sonnet**. Claude is designed to assist in boosting productivity and creativity in various sectors. Some functionalities include:\n",
"- Drafting and debugging code\n",
"- Generating insightful reports and documents\n",
"- Assisting in marketing campaign strategies\n",
"\n",
"### **What Customers Say**\n",
"> “Claude has transformed our workflows, allowing us to accomplish tasks up to **5x faster**!” - Luka Anic, Senior Director, North Highland\n",
"\n",
"---\n",
"\n",
"## Company Culture\n",
"\n",
"**Values that Guide Us**\n",
"- **Mission-Driven:** Focused on ensuring that AI benefits society.\n",
"- **Trust:** We cultivate an open, high-trust environment that allows for honest communication and collaboration.\n",
"- **Teamwork:** Emphasizing collaboration across teams to harness diverse talents and ideas.\n",
"- **Pragmatism:** We value straightforward solutions that effectively balance trade-offs.\n",
"\n",
"---\n",
"\n",
"## Careers at Anthropic\n",
"\n",
"Join our innovative team and make a difference in the AI landscape. We offer:\n",
"- **Competitive Compensation**: Salaries that reflect your expertise, with significant equity options.\n",
"- **Health & Wellness**: Comprehensive insurance benefits, 22 weeks of parental leave, and unlimited PTO.\n",
"- **Flexible Work Environment**: Opportunities for hybrid work and relocation support.\n",
"\n",
"### **How We Hire**\n",
"Our interview process is designed to identify the best talent while minimizing bias. With multiple stages, including exploratory chats and technical assessments, we want to understand not only your skills but also how you align with our mission.\n",
"\n",
"---\n",
"\n",
"## Join Us!\n",
"\n",
"Are you ready to make an impact in the field of AI? Visit our [Careers page](https://www.anthropic.com/careers) to explore current openings and join our mission to develop safer AI solutions for everyone.\n",
"\n",
"---\n",
"\n",
"For more information about Anthropic and our offerings, visit our [Website](https://www.anthropic.com).\n",
"\n",
"**Follow Us:**\n",
"- [Twitter](https://twitter.com/anthropic)\n",
"- [LinkedIn](https://www.linkedin.com/company/anthropic)\n",
"\n",
"---\n",
"\n",
"© 2024 Anthropic PBC | All Rights Reserved \n",
"**Privacy Policy** | **Terms of Service** \n"
],
"text/plain": [
"<IPython.core.display.Markdown object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"create_brochure(\"Anthropic\", \"https://anthropic.com\")"
]
},
{
"cell_type": "markdown",
"id": "61eaaab7-0b47-4b29-82d4-75d474ad8d18",
"metadata": {},
"source": [
"## Finally - a minor improvement\n",
"\n",
"With a small adjustment, we can change this so that the results stream back from OpenAI,\n",
"with the familiar typewriter animation"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "bcb358a4-aa7f-47ec-b2bc-67768783dfe1",
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": 19,
"id": "51db0e49-f261-4137-aabe-92dd601f7725",
"metadata": {},
"outputs": [],
"source": [
"def stream_brochure(company_name, url):\n",
" stream = openai.chat.completions.create(\n",
" model=MODEL,\n",
" messages=[\n",
" {\"role\": \"system\", \"content\": system_prompt},\n",
" {\"role\": \"user\", \"content\": get_brochure_user_prompt(company_name, url)}\n",
" ],\n",
" stream=True\n",
" )\n",
" \n",
" response = \"\"\n",
" display_handle = display(Markdown(\"\"), display_id=True)\n",
" for chunk in stream:\n",
" response += chunk.choices[0].delta.content or ''\n",
" response = response.replace(\"```\",\"\").replace(\"markdown\", \"\")\n",
" update_display(Markdown(response), display_id=display_handle.display_id)"
]
},
{
"cell_type": "code",
"execution_count": 20,
"id": "56bf0ae3-ee9d-4a72-9cd6-edcac67ceb6d",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found links: {'links': [{'type': 'about page', 'url': 'https://anthropic.com/company'}, {'type': 'careers page', 'url': 'https://anthropic.com/careers'}, {'type': 'team page', 'url': 'https://anthropic.com/team'}, {'type': 'research page', 'url': 'https://anthropic.com/research'}, {'type': 'news page', 'url': 'https://anthropic.com/news'}]}\n"
]
},
{
"data": {
"text/markdown": [
"# Anthropic Brochure\n",
"\n",
"**Company Overview** \n",
"Anthropic is a pioneering AI safety and research company based in San Francisco, dedicated to building reliable, interpretable, and steerable AI systems. Our mission is to ensure that transformative AI technologies help society flourish, navigating the complexities and risks associated with the advancement of AI.\n",
"\n",
"**Meet Claude** \n",
"Our flagship product, Claude, is an advanced AI model designed to assist teams and enhance productivity by streamlining workflows and facilitating collaboration. The latest version, Claude 3.5 Sonnet, exemplifies our commitment to safety and performance in AI.\n",
"\n",
"![Claude](link-to-image)\n",
"\n",
"## Our Vision\n",
"At Anthropic, we believe that AI will significantly affect the world. Our aim is to build systems that users can depend on while conducting research into the opportunities and risks AI presents. Our approach to safety is systematic and scientific, integrating research insights into our products while sharing our findings with the wider community.\n",
"\n",
"## Company Culture\n",
"- **Interdisciplinary Collaboration**: Our team comprises experts from various fields, including machine learning, physics, policy, and business. We foster a collaborative environment where diverse perspectives help shape our projects.\n",
"- **High Trust Environment**: We promote an atmosphere of honesty, emotional maturity, and intellectual openness. This trust enhances our decision-making and strengthens our team dynamics.\n",
"- **Mission-Driven Approach**: All team members are aligned with our mission to make AI safer and beneficial for society. We believe in pragmatism, embracing simple yet effective solutions and continuous learning through empirical evidence.\n",
"\n",
"## Customer Focus\n",
"We serve businesses, non-profit organizations, and civil society groups by translating our research into practical tools that facilitate enhanced workflows and better decisions across sectors. Notable sectors benefiting from Claude include engineering, marketing, sales, and customer support, generating measurable improvements in task efficiency.\n",
"\n",
"## Join Our Team\n",
"We seek motivated individuals who are passionate about AI and its societal impacts. At Anthropic, employees enjoy several benefits, including:\n",
"- Competitive salary & equity packages.\n",
"- Comprehensive health, dental, and vision insurance.\n",
"- Unlimited paid time off (PTO).\n",
"- Support for ongoing education and home office improvement.\n",
"- A collaborative and inclusive workplace culture.\n",
"\n",
"### Current Openings\n",
"If you are ready to make an impact in AI safety, explore our open roles on our [Careers Page](link-to-careers).\n",
"\n",
"## Get in Touch\n",
"To learn more about Anthropic and our offerings, visit our website or follow us on our social media channels. We are committed to transparency and eagerly welcome inquiries from interested parties.\n",
"\n",
"- **Website**: [Anthropic](https://www.anthropic.com)\n",
"- **Twitter**: [@Anthropic](https://twitter.com/anthropic)\n",
"- **LinkedIn**: [Anthropic](https://www.linkedin.com/company/anthropic)\n",
"\n",
"### Join Us in Shaping the Future of AI\n",
"Let’s collaborate on building a safer, brighter future with AI technologies. \n",
"\n",
"---\n",
"\n",
"*For press inquiries, please contact: press@anthropic.com* \n",
"*For support and assistance, reach out to: support@anthropic.com* \n",
"\n",
"### Anthropic PBC © 2024\n",
"\n",
"\n",
"---\n",
"\n",
"*This brochure provides a snapshot of Anthropic's mission, culture, and contributions to safer AI technologies. We invite you to explore partnership opportunities and join our dynamic team.*"
],
"text/plain": [
"<IPython.core.display.Markdown object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"stream_brochure(\"Anthropic\", \"https://anthropic.com\")"
]
},
{
"cell_type": "code",
"execution_count": 23,
"id": "fdb3f8d8-a3eb-41c8-b1aa-9f60686a653b",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Found links: {'links': [{'type': 'about page', 'url': 'https://huggingface.co/huggingface'}, {'type': 'careers page', 'url': 'https://apply.workable.com/huggingface/'}, {'type': 'blog', 'url': 'https://huggingface.co/blog'}, {'type': 'company page', 'url': 'https://www.linkedin.com/company/huggingface/'}]}\n"
]
},
{
"data": {
"text/markdown": [
"# Welcome to Hugging Face!\n",
"\n",
"## The Place Where AI Gets *Hugged*!\n",
"\n",
"👋 Hey there, future innovators and trailblazers! Do you want to join a community dedicated to building the future of AI? Then look no further! Welcome to Hugging Face, the vibrant platform where the machine-learning community collaborates on models, datasets, and tons of applications. We promise we won't smother you with hugs... unless you want that!\n",
"\n",
"---\n",
"\n",
"### **About Us:**\n",
"\n",
"At Hugging Face, we're on a heartwarming mission to democratize *good* machine learning, one commit at a time. That’s right, we’re making ML so accessible that even your grandma could model a neural network (with some help, of course). Our fantastic 217-member strong team of AI aficionados is unwaveringly dedicated to making ML fun and fruitful. We even have a secret handshake - just kidding, we don't have secrets or handshakes; we just have **open-source**!\n",
"\n",
"---\n",
"\n",
"### **What We Offer:**\n",
"\n",
"- **Community Collaboration**: Share, discover, and collaborate on **400,000+** ML models and **100,000+** datasets! Our users have gone wild with creativity — from *text generation* to *audio and image solutions!*\n",
"- **Spaces for Everyone**: Like a high-tech playground, our Spaces allow you to run applications like *Kolors Virtual Try-On* and even generate *text-to-video*! Say cheese, please!\n",
"- **AI Tools That Hug You Back**: Dive right into our **HuggingChat**, where you can find vast AI tools available at your fingertips. And don’t worry, they come without awkward small talk!\n",
"\n",
"---\n",
"\n",
"### **Our Customers:**\n",
"\n",
"With more than **50,000 organizations** using our platform, it seems we’re popular! Our clientele includes tech titans like:\n",
"- AI at Meta\n",
"- Amazon Web Services\n",
"- Google\n",
"- Intel\n",
"- Microsoft\n",
"- Grammarly\n",
"\n",
"If you’re looking to join an entourage of heavyweight entities, you’ve landed in the right hug!\n",
"\n",
"---\n",
"\n",
"### **Careers:**\n",
"\n",
"🎉 **Join the Hugging Face Family!** 🎉\n",
"Our door is always open for creative minds who want to sprinkle a little magic into the world of AI. We're on the hunt for talent in various areas. Think you can help us? Here's what you can expect:\n",
"- A team that celebrates diverse perspectives — we embrace differences like we embrace our cats!\n",
"- A chance to build your ML portfolio & get recognized (we mean MORE than just a thumbs-up emoji!).\n",
"- A flexible work environment where you can wear slippers to meetings if you choose (we won’t judge).\n",
"\n",
"---\n",
"\n",
"### **Closing Thoughts:**\n",
"\n",
"So there you have it! At Hugging Face, we’re committed to building a friendly, collaborative, and innovative environment. Whether you're a customer, investor, or a potential recruit, remember: **We believe in hugging, not hacking!**\n",
"\n",
"> So what are you waiting for? Join us, and let's build the future together! 🚀💚\n",
"\n",
"[Visit Hugging Face](https://huggingface.co) to become part of our connected community! "
],
"text/plain": [
"<IPython.core.display.Markdown object>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"stream_brochure(\"HuggingFace\", \"https://huggingface.co\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "bcf5168e-f1d9-4fa7-b372-daf16358e93c",
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.11.9"
}
},
"nbformat": 4,
"nbformat_minor": 5
}