Get a reply from this employer
Your resume and cover letter, tailored to exactly what they’re hiring for.
Unloq is hiring an outstanding AI Engineer to own the quality loop on our decision engine. You will verify each client’s answers against ground truth, ensure changes don’t degrade performance, and contribute to end-to-end delivery from data processing to answer presentation.
You should bring 1–3 years post-internship experience building AI or backend systems, strong Python and SQL, and hands-on exposure to LLM/RAG/text-to-SQL projects. Remote work with IST core hours is offered.
This is a very exciting time for companies at the frontier of AI, and Unloq is scaling fast. We’re building the system of record for business decisions, held to one standard: every answer has to be defensible to a board. We’re VC-backed, going live with important enterprise customers, and building towards our Series A.
That makes this a rare chance to join a startup early in its growth, while the team is small, the big problems are still open, and the people who join now shape what Unloq becomes.
We’re looking for an outstanding AI Engineer to join our platform team and own the quality loop on our decision engine: proving each new client’s answers against ground truth before they go live, and making sure no change makes them worse. You’ll report to our Founding Lead Engineer for the decision engine and work day to day with our founding leads. Most of your week is code, experiments and evaluation runs. The rest is code review, benchmark reviews and pairing.
Building and verifying each new client’s golden set: real questions and the correct SQL, checked against ground truth before go-live
Running every change through the benchmark, and turning every thumbs down into a test, so nothing gets worse
Making text-to-SQL repeatable, so a re-run gives the same SQL and scores only move when the code does
Shipping across the engine, from document search and the Semantic Layer to how answers are shown, holding back any chart or table we can’t verify
you’ll ship your first fix, take a new client’s golden set to go-live, and turn their thumbs downs into tests.
1 to 3 years beyond internships, building AI or backend systems
Strong Python and SQL. Non-negotiable: you can spot a query that runs cleanly and returns the wrong answer
Something real. At least one LLM, RAG or text-to-SQL project, with a way of measuring whether it worked
Honest numbers. You’d rather report an honest 70% than a flattering 95%
High agency. You take a loosely defined problem and drive it to done
AI-native. Claude Code daily, and you own every line you merge
a high GPA B.Tech from an IIT, IIIT, NIT or BITS, evaluation or forecasting work, a Kaggle ranking, hackathon win or open-source contribution.
A query returns the right columns but sums revenue including returns. How would you catch it automatically?
Your evaluation score jumped overnight with no code change. What happened?
Location: Remote, 12pm to 9pm IST core hours
Start: ASAP
Process: a conversation, a short paid SQL exercise, a technical round, and a final with our CTO
Learn from the builders. Work day to day with the founding engineers behind the engine.
Real ownership. Your benchmark decides what ships.
Modern stack. Claude Code, Python, FastAPI, PostgreSQL, Docker, AWS and Azure, Databricks, and Langfuse for evals and tracing.
If you’d rather catch a wrong answer than ship a confident one, this is your role.