💻 AI Architect @ BLUE — building a real-time multi-camera VLM pipeline for warehouse and retail vision analytics.
💻 Former Founding AI Engineer @ Stealth — designed a 27-agent orchestration framework and multi-stage document enrichment pipeline for AI-powered factory floor planning.
💻 Former Senior AI Engineer at Avathon — scaled a CV platform to 600+ cameras across enterprise deployments, and introduced LLM + VLM layers to a purely CV stack.
💻 Former Deep Learning Engineer at TCS — shipped CV models for Smart Mobility and automotive data pipelines.
💻 Former Intern at Regional Remote Sensing Centre, ISRO.
🎓 B.Tech in Computer Science & Engineering, 2021.
🌱 Working across LLMs, VLMs, multi-agent systems, and on-device inference.
🤗 Publishing open-source MLX-quantized LLMs on HuggingFace — making frontier models runnable locally on Apple Silicon.
💬 Feel free to reach out for projects, collabs, or just an interesting discussion.
- Comparative study — Using EfficientNetB0 to EfficientNetB7 as encoder in UNet
- Vanilla UNet vs UNet with EfficientNetB0 as Encoder
- Setting up Nvidia’s Jetson Nano : From Jetpack to YoloV5
- Run YoloV5s with TensorRT and DeepStream on Nvidia Jetson Nano
- Paper summary — Decoupled Weight Decay Regularization
- Paper Summary — What is being transferred in transfer learning?
- Paper Summary — torch.manual_seed(3407) is all you need
- Paper Summary — MetaFormer is Actually What You Need for Vision
- Dive into basics of GPU, CUDA & Accelerated programming using Numba in Python
- Video Analytics Dashboard for YoloV5 and DeepSort
- Intel OpenVINO in brief
- All you want, to get started with GStreamer in Python
- Paper Summary — Searching for TrioNet
- Paper Summary — Efficient Long Range Attention Network (ELAN)
- Paper Summary — RepVGG : Making VGG-style ConvNets Great Again
- Paper Summary — ShuffleNetv2 : Practical Guidelines for Efficient CNN Architecture Design
- Paper Summary — Deep Leakage from Gradients
- Common questions while working with Large Language Models
- Merged > Fine-Tuned? A Case Study on Qwen3 and Domain Fusion
- Fine-Tuning LLMs for Refusal
