Krish Ujeniya

Meet Krish

1 following

I build production AI systems that actually ship—typically in under four weeks.

Most ML builds get stuck in notebooks or break quietly in production. I treat every project like core infrastructure: with evaluation suites, error fallbacks, drift monitoring, and clean documentation so your team can maintain it without friction.

Core capabilities:
• Enterprise RAG: Hybrid search, reranking, and automated evaluation pipelines (RAGAS) to surface exact answers from company documents.
• Autonomous AI Agents: Multi-agent systems with LangGraph and CrewAI, featuring structured memory, deterministic tool calling, and human-in-the-loop controls.
• MLOps & Delivery: Reproducible training and inference workflows with ZenML, MLflow, and cloud deployments.

Selected builds:
• Hadron — Autonomous GitHub PR review agent using LangGraph and hybrid vector retrieval.
• AmberClaw — Lightweight conversational memory engine maintaining cross-session user context.
• SnapText — On-device OCR mobile application running real-time text extraction with zero cloud dependencies.

Background:
Mathematics and software engineering. Certified across Azure AI and cloud infrastructure with practical experience shipping ML systems for startups and product teams.

Working together:
Weekly demos, direct communication, and full code handover on day one. Available for contract engineering, system architecture, and production builds worldwide.