Jyotirmoy Konwar
Research Intern @ Siemens | PyTorch, Quantization, MLOps, LLMs | Top 52/9k+ Amazon ML Challenge 2025
- Role
- Research Intern at Siemens
- Location
- Bengaluru, KA, IN
- LinkedIn followers
- 500 followers
About Jyotirmoy Konwar
I’m an MLOps & AI Engineer with 2+ years of hands-on experience productionizing multimodal and LLM systems that actually ship to production. At Borde (MLOps Intern), I cleaned and preprocessed a 120K+ image dataset and fine-tuned, deployed and monitored YOLOv5 models on AWS and Azure GPU instances — delivering 60% overall accuracy improvement and 150% precision lift on low-performing labels. Most recently in my freelance project, I built a zero-shot multimodal parsing pipeline with Qwen 2.5 VL + 4-bit GGUF quantization (llama.cpp), achieving 98% extraction accuracy (↑16% from baseline), 8× memory reduction, and full CPU-only inference — processing ZIP archives in batch and exporting structured Excel outputs. Along with these, I have had multiple research internship opportunities. On the side, I architected a 3-stage SFT + GRPO fine-tuning pipeline for Gemma-3 (12% gain in answer accuracy and 47% surge in reasoning quality), trained a Tiny Diffusion Model from scratch that tells you tiny stories and ranked #52 out of teams in the Amazon ML Challenge 2025 with a multimodal pricing ensemble model using DeBERTa, ELECTRA, and CLIP-ViT. I specialize in taking research models and turning them into reliable, cost-efficient production systems — LoRA adapters, quantization, batch inference, monitoring, and end-to-end MLOps. Passionate about efficient GenAI inference, multimodal systems, and real business impact. Currently open to ML Engineer, AI Engineer, ML Systems Engineer, or Applied Scientist roles in Bengaluru or remote. If you’re building production AI, let’s connect — I’d love to share my GitHub repos or discuss how I can contribute to your team! 🤝
Experience
Research Intern
May 2026 — Present · Bengaluru, IN
Education
Rajiv Gandhi Institute of Petroleum Technology (RGIPT)
Integrated Dual Degree in Computer Science(B.Tech) and Artificial Intelligence(M.Tech), Computer Science
Sai Vikash Vidya Niketan
JEE rank- 50,048 JEE Adv rank- 19k (Qualified) Assam CEE rank- 64
2020 — 2022
Find verified contacts for anyone on LinkedIn
Unifers gives sales teams verified emails and direct dials, enriched profiles, and outreach that lands in the inbox.
Free plan included · No credit card required
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.