AI Agent Architect

Overview

AI Agent Architect to help build AI agents that plan, build, test, and ship production software from plain-language intent, improving measurable performance end-to-end.

What you'll do

• Turn LLM and system capabilities into measurable, shippable improvements in agent performance.
• Own the improvement loop end-to-end: what to try, how to prove it works, what goes live, and what gets rolled back.
• Mine production behavior to identify failure modes, regressions, and bottlenecks and convert them into quantitative signals.
• Define and run experiments to improve agent quality, reliability, and code outcomes across prompt tuning, eval datasets, and harness engineering.
• Build evaluation frameworks to measure agent quality at scale and ship dashboards for regression visibility.
• Use metrics, evaluation gates, staged rollouts, and rollback criteria to ship with rigor.
• Work on context engineering, memory systems, tool use, and long-horizon execution.
• Make decisions in probabilistic systems about when regressions are real, wins are noise, and when to ship or stop.

What you'll need

• 5+ years building and shipping software with real end-to-end ownership.
• At least 1 year of hands-on experience with agentic systems and their evaluation.
• Strong systems thinking with sharp product judgment.
• Hands-on experimentation, metrics, and debugging.
• Fluency in Python, SQL, or similar for analysis.
• Comfort reasoning about noise, confounds, distribution shift, and selection effects.
• Deep, current interest in LLMs, agents, and emerging model capabilities.

Details

• Location: Bengaluru.