AI powered people search
The Role: Software Engineering Intern
As an engineering intern you'll have a hands-on role in building PeopleGPT, and will be treated like a full-time team member. You will work closely with the founders, Ishan and David, as well as the engineering team, on technical projects ranging from iterating on LLM chains to building frontend components and optimizing search performance.
You will be onboarded in the same way our full-time team members are on boarded --- and instead of working on intern projects that never make it to production, you'll be shipping code from week one.
You will design, build, and execute product features from end to end --- many of which will be built upon LLMs. Projects may include things like:
Implementing new LLM-powered filters, such as "find similar companies to Tesla"
Evaluating performance of open-source LLM models on 100k+ datapoints
Setting up an inference pipeline to run GPT3.5 on 800mm profiles
Building and running a matching algorithm to match job descriptions with security clearance requirements
Developing an end-to-end workflow for a "clone this profile" search
You will engage with customers, investors, and the founding team to develop your own theses on product improvements, directions, and initiatives.
Our current tech stack includes:
Next.js + React in the frontend
Node.js in the backend
Firebase, OpenSearch, Kubernetes, and several AWS + GCP services
Open Source + OpenAI LLM Models
About You
You're hungry to build something huge.
You have experience working/interning at a startup, or in a lean, high-performance team.
You've shown excellence: be it in your education, work experiences, projects, or anything else you can share.
You take asymmetrical bets.
You are hands-on, fast and learn best by building.
You have experience with the core of our stack: Next.js, React, Firebase, AWS, GCP
You are keen to become an expert in search, vector embeddings and ML.
Bonus:
You've built features with LLMs or launched a project in space.
You want to take on a leadership role in the future.
You have experience in the with search infrastructure (OpenSearch, vector search)
Our Values
Hyper focused on the customer.
Speed wins: ship fast.
Small teams enable outsized impact.
Presence matters. In office, 5 days a week.
You own the business. We are generous with equity grants.
Transparency. We share what is going well, and what isn’t.
Iteration. Things rarely work on the first try. We keep shipping.
First Interview: with David Paffenholz. You'll speak about your motivations, experience, and have the opportunity to ask questions about the role. Expect 15-30 mins.
Second Interview: with Ishan Gupta. You'll dive deeper into your experience with Juicebox's tech stack (15 mins) followed by a technical section (30 mins). The technical section involves questions around fundamental programming concepts, your experience around building large-scale / distributed systems, and some of the relevant frameworks you already have experience with. Expect 45-60 mins.
Third Interview: with Ishan Gupta. This interview has two sections: a technical assessment (30 mins) followed by a system design section (30 mins). For the technical assessment, you will focus on data structures, algorithms, graphs, and hands-on skills with JS web frameworks. The system design section will assess your ability to design large-scale systems. It will not involve a coding task. Expect 45-60 mins.
Fourth Interview: with David Paffenholz. You'll discuss a product case study, in which you will dive deeper into potential user pain points and iterate on solutions to them (20-30 mins). Next, you will talk about your future goals and ambitions. Expect 30-45 mins.
All business processes start with searching for people: the most talented person to hire, your next promising sales lead, or the expert who can solve your engineering problems. For a long time, the people search industry has been a monopoly.
But AI has created an opportunity to disrupt it.
Juicebox is an AI-powered people search engine. We enable search on top of 800+ million profiles using natural language. We have over 25,000 users, including hundreds of paying customers ranging from public companies to fast-growing startups.