Pranav TJ

I've built distributed maritime systems at ZVIA Technologies and full-stack EV charging products at NexEV Solutions, delivering production applications across AWS and Azure.

At ZVIA, I architected a Dockerized FastAPI and PostgreSQL platform with multi-tenant isolation, JWT-based RBAC, replication, and 100+ endpoints. I also reduced LLM response latency by 35% and query resolution time by 40% through a fine-tuned Llama3 and contradiction-aware RAG pipeline.

At Parsewave, I built multimodal training and evaluation data, RL environments, and AI orchestration workflows that contributed to an 8-point improvement in Meta's Muse Spark 1.1 performance on multimodal SWE benchmarks.

I'm equally hands-on with AI products and core systems, from an SSH honeypot with immutable Hyperledger Fabric logging to a MERN and FastAPI learning platform serving 200+ students.