Jyotiranjan panda

I've spent ~19 years in tech, the last several focused on the unglamorous part of

AI: getting systems past the demo and into production where they have to survive

real traffic, real compliance, and real cost ceilings.

Recently I architected a HIPAA-compliant document intelligence pipeline

processing 10,000+ PDFs daily at 92% extraction accuracy, cutting 200+ manual

hours a day, and built an FDCPA-compliant outbound voice agent with sub-2s

response latency and policy enforcement in the loop. Across enterprise services

and communications-platform roles, I've converted 10+ AI proofs of concept into

production systems and supported $15M+ in AI/ML presales. Earlier in my career I

scaled a C/C++ messaging platform to 20 billion messages a day at 99.99% uptime —

that systems background is why my AI work tends to hold up under load.

I work hands-on across multi-agent orchestration, RAG and retrieval pipelines,

LLMOps and evaluation, real-time voice (STT/TTS/barge-in), and vision-language

document extraction — mostly Python on AWS. I hold a US patent for deep-learning

automation, have published on arXiv, and am AWS Certified Generative AI

Developer – Professional.

I'm open to remote senior individual-contributor or leadership roles with teams

building production GenAI systems, not pilots.