Founding Engineer (Data + AI)

About this search

Opentap Talent is recruiting a Founding Engineer on behalf of a stealth-mode consumer AI startup building for the Indian market. Our client is past concept stage with a live MVP, real users, and active institutional interest. Well share the company name and full context with candidates who reach the conversation stage.

The opportunity in 30 seconds

Our client is building an AI + Human powered career advisor for Indian students. Not another job board. Not another quiz. A genuinely intelligent product that builds personalized, dynamic career plans and stays with users as their journey evolves. The base version of the product is live as an MVP. Early users include students from real Indian colleges.

What youd own

• The data architecture. Every interaction in this product produces high-signal data: who the user is, what was recommended, what they did, what happened next. Designing how this data flows, gets labeled, and compounds into a defensible moat is the core competitive advantage. Youd own schema design, labeling discipline, retrieval patterns, and feedback loops.
• The AI generation pipeline - specific factual claims (companies, salaries, timelines) must be grounded in a verifiable data layer to prevent fabrication. Designing this three-layer pipeline (LLM path discovery factual grounding personalized narrative wrapper) is non-trivial and central to the product.
• The learning loop. As the product scales, it has to get smarter. That means building inference logs that capture model decisions, outcome tracking that closes the loop on whether recommendations worked, and retrieval systems that let future inferences pull from past patterns. Youd own this from scratch.
• The infrastructure. Supabase, Claude API, edge functions, vector storage (or alternatives), and an embedded analytics layer. Whatever it takes to ship cleanly and at quality.

What our client is looking for

• Strong product instincts. Not someone waiting for a spec. Someone who pushes back on product decisions, asks the right architecture questions before they become problems, and brings a point of view.
• 3+ years building data-heavy or AI-powered consumer products. Bonus if youve shipped something with retrieval-augmented generation, structured + unstructured data hybrid systems, or anything where output quality depends on careful prompt-data interactions.
• Comfort with early-stage ambiguity. Things will shift. Frameworks will change. Theres no head of engineering filtering decisions. Youd be making them with the founder.
• Indian context. Youve lived in India, built for Indian users, or have a clear instinct for what makes a product feel built-for-Bharat rather than copy-pasted from Silicon Valley.
• Honesty over politeness. The founder pushes back when they disagree. They want someone who does the same.

Compensation

Equity: Real founding-engineer range (1-3%, vesting standard)

Cash: Negotiable based on your situation. Realistic for an Indian founding-engineer market; the founder is open to discussing what makes this work for the right person.

Ownership: Employee #1 on the technical side. Your fingerprints on every architectural decision.

Direct work with the founder. No middle layer.

First 60 days look like

Pairing on data architecture (currently mid-design)

Shipping the inference log schema and outcome-tracking layer

Architecting the LLM path generation pipeline with verifiable grounding

Setting the foundation for v1.5: success-journey dataset, retrieval systems, broader market intelligence

After that, youd grow into a more permanent role as the technical co-builder.

Shoot your resume and your answer to the following question to info@opentaptalent.com to skip the queue.

When designing the data layer for an AI product that needs to get smarter over time, what schema decisions and labeling discipline matter most from Day 1, and why?