AI / ML Engineer (Unpaid for OPT Credit)
AI / ML Engineer (Unpaid for OPT Credit)
Posted 2 days ago
SENIORITY
Senior
About the role
- A recent graduate and AI/ML engineer with experience shipping production-grade LLM applications that serve real users, not just prototypes or toy examples.
- Experienced in building LLM-powered products end to end, including retrieval-augmented generation (RAG), agent and tool-calling workflows, and the evaluation systems that keep them reliable.
- Strong in Python and comfortable owning services from design through deployment, monitoring, and iteration.
- Excited about conversational intake experiences, and about making AI systems accurate, safe, and trustworthy in a high-stakes legal setting.
- Curious about retrieval, evaluation, and how people find the right attorney, even if you don’t have prior legal domain experience.
- Energized by working in a fast-paced, early-stage startup where you’ll wear multiple hats, navigate ambiguity, and make decisions quickly.
- Our Tech StackYou don’t need to know everything on day one, but you should be excited to work with (and learn) most of this stack:
- Backend & APIs: Python, FastAPI, Uvicorn, and Pydantic.
- Frontend & UX: React and TypeScript (built with Vite and Tailwind CSS).Orchestration & Agents: CrewAI for multi-agent orchestration on Link, and NVIDIA NeMo Guardrails for conversational safety.
- AI / LLM Providers & Libraries: OpenAI GPT models, Google Gemini, and LangChain with LangSmith for tracing and evaluation.
- Messaging: Twilio for phone verification.
- Data & Storage: PostgreSQL on Google Cloud SQL.CRM & Business Integrations: Salesforce CRM (REST API integration, Apex, and Lightning Web Components), and Stripe billing.
- Cloud & Infrastructure: Google Cloud (GCP), including Cloud Run, Cloud Build, Artifact Registry, Secret Manager, Cloud Tasks, and Cloud Storage, plus Docker and GitHub / GitHub Actions for CI and security scanning.
- Analytics & Logging: PostHog product analytics and Google Cloud Logging to understand usage and model behavior.
- Experience building conversational AI experiences.
- Experience with Google Cloud Platform services such as Cloud Run, Cloud Build, and Cloud SQL.Experience integrating with Salesforce or other CRMs, or with legal practice-management tools.
- Experience with Twilio or other messaging and phone verification APIs, and with Stripe or other billing integrations.
- Familiarity with application security practices and tooling, such as Bandit, Semgrep, Truffle
- Hog, and CodeQL.Familiarity with TypeScript and React (with Vite and Tailwind CSS) so you can collaborate closely with frontend engineers and designers.
- Experience with product analytics tools such as PostHog to connect model changes to user outcomes.
- Background in NLP, ranking, or recommendation systems, which power how we match cases with the right attorneys.
- Interest in the legal industry and in making it easier for people to find the right attorney.
- Build and improve the AI that powers Link, our consumer intake and attorney matching product in, from conversational intake through matching consumers with attorneys who have handled similar cases.
- Improve the matching and ranking logic that connects people with attorneys who have handled similar cases.
- Design and improve RAG pipelines: chunking, retrieval, ranking, and context assembly.
- Build and orchestrate agents and tool-calling workflows, with guardrails that keep conversations safe and on task.
- Prototype to Production:
- Move quickly from experiments and prototypes to reliable, maintainable production services on Cloud Run.
- Own testing, monitoring, and deployment for the systems you build, and help raise the team’s engineering standards.
- Data & Ingestion Work:
- Build ingestion and processing for the data behind intake and attorney matching, including case records, intake conversations, and other unstructured documents.
- Design schemas, embeddings, and features that make retrieval and reasoning more accurate and robust.
- Evaluation & Experimentation:
- Design and run evaluations that measure model quality, retrieval performance, and user-facing outcomes.
- Build evaluation suites and dashboards that track quality over time (e.g., accuracy, groundedness, hallucination rate), and use product analytics to connect model changes to user outcomes.
- Cross-Functional Collaboration:
- Work with product, design, and legal experts to understand user needs and turn them into concrete technical plans.
- Participate in sprint planning, code reviews, and technical discussions as part of a small, high-ownership team.
- Compensation
Before you apply
Applying takes about a minute. These four things decide how fast it moves after that.
Your profile is current
It's what we read first. Occupations, seniority and locations matter more than a long history.
Two examples you can talk through
Not a portfolio — just two pieces of work where you can explain the decisions and what you'd change.
A number in mind
What you're on now and what would make you move. We negotiate better when we know both.
Your notice period
Employers plan around it, and it's the question that stalls offers most often.
Once you apply, someone reads it and calls you before anything reaches the employer — usually within two working days.
