15+ years engineering, last 4 deep in production LLM systems and agentic AI infrastructure.
Recent work: LLM fine-tuning pipelines, MCP server ecosystems, RAG integrations, and AI tooling
for large-scale developer documentation — the kind of problems where "it must work reliably in
production" is the hard constraint.
A few things I've shipped:
First ML model for Grade 2 contracted Braille OCR — the format used in 90%+ of real documents, unsupported by any existing OCR. ByT5 seq2seq on 25K synthetic pairs; 92.9% exact match on real-world Braille, ~9× better than any prior baseline.
https://huggingface.co/prasanthmj/braille-byt5-v3
Looking for senior/staff AI engineering roles — particularly where LLM systems, AI governance, or applied ML research intersect with real production constraints.
Recent work: LLM fine-tuning pipelines, MCP server ecosystems, RAG integrations, and AI tooling for large-scale developer documentation — the kind of problems where "it must work reliably in production" is the hard constraint.
A few things I've shipped: First ML model for Grade 2 contracted Braille OCR — the format used in 90%+ of real documents, unsupported by any existing OCR. ByT5 seq2seq on 25K synthetic pairs; 92.9% exact match on real-world Braille, ~9× better than any prior baseline. https://huggingface.co/prasanthmj/braille-byt5-v3
gomcpgo — Go MCP framework + 12 open-source servers. Single binary, zero runtime deps. https://github.com/orgs/gomcpgo/repositories
Looking for senior/staff AI engineering roles — particularly where LLM systems, AI governance, or applied ML research intersect with real production constraints.
LinkedIn: https://linkedin.com/in/prasanthmj