Boston, MA · open to full-time
AI Agent Engineer in Boston
Abhishek Sagar Sanda is an AI agent engineer building voice agents, retrieval systems, and the full-stack product that keeps them reliable in production - not only in a notebook.
What an AI agent engineer ships
An AI agent engineer designs systems that act under real constraints: live phone calls, latency budgets near 800 milliseconds, tool use, and failover when a model or network path fails. That means owning telephony (Twilio, LiveKit), orchestration, evaluation, and the product surface callers or users actually touch.
As an AI agent engineer, Abhishek focuses on two production patterns: conversational voice agents that answer inbound calls end to end, and retrieval-augmented agents that cite sources or refuse when the corpus does not support an answer. Both demand measurable loops - not demos that only work on a laptop.
Proof from shipped systems
- BitVoice Pay - voice-first Lightning wallet over a phone call; live dial-in proof of agent behavior on Twilio.
- Northeastern University Assistant - RAG over 80,000+ pages with citations, the retrieval half of agent grounding.
- AI Interview Coach IVR - multi-turn interview coaching on a live call path.
- AutoAce (YC F25) - voice-agent OS for dealerships: latency cut 8.5s → 3.9s, backfilled sessions, three-layer LiveKit fence.
Hire an AI agent engineer
Based in Boston and available for full-time roles, Abhishek works with Series A–C teams that need an AI agent engineer to take voice and retrieval products from prototype to dependable service. Related: forward deployed engineer, full-stack AI engineer. Start with the selected work, read the resume, or book a call.