I'm an engineer who started out in electronics, fell for signal and image processing, and ended up spending two years teaching neural networks to find people in thermal images taken from drones. These days I build applied AI systems at Ascendion in Bengaluru.

What I care about most is understanding why a model works, not only that it does. I like small, efficient architectures, honest evaluations, and measuring what matters instead of what's easy to measure. Outside of work I read about cosmology and spacetime, which is a nice reminder that there are harder problems than class imbalance.

This site is where I write down what I learn. Some of it is polished, a lot of it is notes. If something here was useful, or wrong, I'd love to hear about it.

Contact

What I'm doing now

See the now page, updated September 2026.

Experience

  1. 2026โ€“now

    Ascendion ยท Bengaluru

    Engineer, applied AI

    Building applied AI systems for clients: retrieval pipelines, LLM integrations and agentic workflows.

  2. 2025โ€“2026

    PRISM Research Group, NIT Rourkela

    PG researcher with Dr. Sobhan Kanti Dhara

    Designed and led four lightweight detection architectures (TSRNet, SAPNet, CGANet, C2SNet) for tiny objects in aerial thermal imagery, and co-authored work on ship detection. See the papers โ†’

  3. 2022

    Space Applications Centre, ISRO ยท Ahmedabad

    Research intern with Dr. Deepak Putrevu and Dr. Anand S. Sahadevan

    Processed large airborne LiDAR datasets of forest canopies and wrote a region-growing algorithm to separate individual tree crowns, then checked it against ground truth.

Education

  1. 2024โ€“2026

    National Institute of Technology, Rourkela

    M.Tech, Signal and Image Processing

    Thesis: Deep Learning Architectures for Tiny Human Detection in Aerial Thermal Imagery. IndiaAI Fellowship recipient.

  2. 2019โ€“2023

    Jabalpur Engineering College

    B.Tech, Electronics and Communication Engineering

Toolbox

Languages
Python, SQL, MATLAB, C
Deep learning
PyTorch, MMDetection, Ultralytics YOLO, TensorFlow
Applied AI
RAG pipelines, FAISS, LLM APIs, prompt engineering
Deployment
ONNX, TensorRT, Docker, edge inference
Favourite topics
Object detection, efficient architectures, thermal imaging, model evaluation

Elsewhere on this site

  • Research: papers with plain-language summaries
  • Projects: things I've built
  • Blog: notes and longer posts
  • Photos: the non-work side