Utsav Akhaury

 

prof_pic.jpg

I’m an engineer with a long-standing passion for AI/ML and computer vision — one I turned into a PhD.

I now apply that foundation as an AI Software Engineer at Spot AI, working on Agentic Vision — AI that sees, reasons, and acts on real-time video.

I hold a PhD in Physics from École Polytechnique Fédérale de Lausanne (EPFL). My thesis focused on deep learning–based image enhancement, working with data from flagship NASA and European Space Agency (ESA) missions like Hubble, James Webb, and Euclid.

Across 5+ years, I’ve built, trained, and shipped ML systems from prototype to production, with deep experience in HPC environments (multi-node GPU training, distributed data pipelines). In 2021, I was a Google Summer of Code contributor, building ML surrogates for physics-based simulation.

What I work on

Perception Detection, tracking, and segmentation for real-time video (YOLO, ByteTrack, SAM 2)
Restoration Denoising, deconvolution, and super-resolution for astronomical imaging (U-Net, ViT, Swin Transformer)
Multi-modality VLMs and CLIP-based retrieval for natural-language video search

Tech I work with

Languages Python C++ Julia MATLAB Rust
ML / Deep Learning PyTorch CUDA Hugging Face TensorFlow scikit-learn
Computer Vision OpenCV ONNX TensorRT FFmpeg
MLOps & Tooling Slurm Docker Kubernetes Google Cloud GitHub Actions

Outside work

My journey has taken me across India, Mauritius, France, and Switzerland, making me quietly obsessed with linguistics.

Along the way, I’ve developed a creative side as a pianist and electronic music producer. I’ve been playing piano since the age of 8 and have performed live at venues ranging from Parisian bars to cultural events at the United Nations, Geneva. Check out my compositions and covers on the music page.

News

Mar 24, 2026 Joined Spot AI as an AI Software Engineer 🎉
Feb 12, 2026 Defended my PhD at EPFL — Dr. status unlocked!
Sep 01, 2021 Successfully completed Google Summer of Code