AI-Native Product Engineer

Company Description & Culture Code

https://optima.ai/talent

The Idea

Models are a commodity. The harness is the product. The evolution loop is the moat. Anyone can call an LLM; the difference between an impressive demo and a product clients trust with their money lives in the harness (context engineering, memory architecture, tool design, runtime discipline, generative UI, observability) and in the evolution loop that tightens it over time. Optima has to feel at least as good as the best AI products people use today; Claude is the reference bar, not a lofty aspiration. That bar is not cleared with prompts. It is cleared with engineering. We need one head holding that foundation: systems-proficient enough hold production load, AI-native enough to think in agents and traces, versatile enough to ship anything else the work demands. If you know how to build the foundation a state-of-the-art AI product runs on, we're talking about you.

Role Description

We are recruiting an AI-Native Product Engineer to own the harness underneath Optima's AI product surfaces:

• Own the harness end-to-end: substrate (context engineering, memory architecture, tool design), runtime (determinism budget, single-agent defaults), interface (bidirectional context, typed generative UI, narrative sequencing), and observability (trace pipelines, evals, A/B infrastructure)
• Treat product design as a first-class craft: language, interaction feel, narrative pacing, systems thinking are not just decoration
• Ship versatile when the work demands it: when a product surface needs systems backbone, internal tooling needs scaffolding, or any other engineering work has to land, you do it (or build the system that does)
• Raise the floor for everyone around you: your code, your reviews, and your architectural choices set the bar that the rest of engineering pulls toward

Who This Is For

• You have 4+ years shipping production systems end-to-end at product-driven companies where craft was the competitive edge
• You hold strong academic signal from a premier CS program (IIT CSE / IIT MnC or equivalent)
• You find finance interesting enough that regulatory complexity and operational mess read as constraints to absorb on behalf of the user, not friction to push downstream
• You think in traces, latencies, error budgets, and architectural primitives, not in features and tickets
• Or something untraditional that trumps all of the above

What This Isn't

• Not a research role
• Not a platform or devops role
• Not a pure ML role
• Not a people management role

Compensation Highlights

• Top-of-market compensation (₹50L - ₹1Cr PA)
• Health insurance

Contact

Share something you've built (a product, a demo, or a technical write-up/architecture diagram of a hard problem you solved) directly with us at [Confidential Information].