Ishaan Katara

About Ishaan Katara

I’m an ML Systems Engineer with a strong track record of turning research models into production-grade features, combining deep learning (PyTorch, Transformers), scalable backend engineering (Python, FastAPI, TypeScript, Node.js, NestJS), and cloud-native DevOps/DevSecOps (Docker, Kubernetes, CI/CD, AWS). At Cinntra I helped design and deploy RAG pipelines and fine-tuned LLaMA-family models, delivering measurable gains (e.g,~18% domain accuracy improvement from targeted fine-tuning,~38% inference cost reduction via quantization/batching, and a 35% faster training-to-deploy iteration through MLOps automation). I’ve built production RAG services that handle 10K+ document chunks with p50 latency ≈320 ms and p95 ≈1.8 s while keeping retrieval relevance high, and I excel at integrating ML into product flows using vector search (FAISS/pgvector), orchestration (Kafka, Redis), and monitoring/observability. I’m actively looking for ML/AI Systems and Backend engineering roles where I can scale ML-powered products end-to-end and drive measurable business impact.

Experience

1. Founder

   Stealth Startup

   Jan 2026 — Present · New Delhi, IN

Education

• Jaypee Institute Of Information Technology

  Bachelor of Technology - BTech, Computer Software Engineering