🚗📡 Multi-modal perception, 3D reconstruction, and generative deep learning — engineered for the real world.

I design and build systems that make sense of the world by fusing data from LiDAR, radar, cameras, GPS, and IMUs—turning raw signals into robust 3D understanding.

My work spans from360° automotive auto-labeling frameworks and3D reconstruction pipelines togenerative modeling withdiffusion andflow-matching techniques.

I move comfortably between deep learning and classical algorithms, and across many fields within AI—from segmentation and 3D object detection to vision-language and multi-modal fusion.The aim is always the same:engineer solutions that work in the real world.

Before I specialized in AI, I built my foundation in hardware, embedded systems, and high-speed digital design—an early career chapter that shaped my engineering mindset and my obsession with building things that last.

I don’t just run models; I dig into them. Ifine-tune foundation models, adapt them for specific sensor setups, and optimize them to run faster and smarter.
Each project I take on pushes my skills further—my latest repos are always my best yet.

• Code accompanying my research papers, published in venues like IEEE Access and arXiv
• Dotfiles, setup scripts, and tools I use to optimize my Linux dev environment
• My personal website and related static-site tooling
• Occasional experiments and utilities from my engineering workflow

Most of my day-to-day AI work happens in industry and stays behind closed doors — but you can find more of my thinking and projects here:

• 📄LinkedIn — career highlights, leadership work, and public talks
• 📝Medium — deep dives into AI concepts, research walkthroughs, and engineering insights

If you care about wringing every drop of performance from a model and fusing diverse data into something meaningful, we’ll get along.

PinnedLoading

1. Communication_ModulationCommunication_ModulationPublic

   This repository provides a simple python script for getting experience with common modulation techniques i.e. QAM, PSK, ASK and BPSK

   Python 95 20

2. ML_NotebooksML_NotebooksPublic

   Collection of machine learning related notebooks to share.

   Jupyter Notebook 19

3. VoloGANVoloGANPublic

   Official code for the paper "VoloGAN: Adversarial Domain Adaptation for Synthetic Depth Data"

   Jupyter Notebook 2 1

4. rgb-d-fusionrgb-d-fusionPublic

   Official implementation of the paper "RGB-D-Fusion: Image Conditioned Depth Diffusion of Humanoid Subjects"

   Python 9 3

5. samba-mixersamba-mixerPublic

   Official implementation of the paper "SambaMixer: State of Health Prediction of Li-ion Batteries using Mamba State Space Models"

   Python 24 2

6. linux-forgeuplinux-forgeupPublic

   Forge-Up is a lightweight setup and bootstrap tool for Debian/Ubuntu-based systems.

   Shell 2