Akshay Gautam

AI engineer working across models, agents, vision, and robotics.

I work across the practical AI stack: collecting and cleaning data, creating datasets, fine-tuning models, building retrieval systems, and making agents usable in real products. I have worked professionally in software, and I also build robots and hardware systems on the side.

X GitHub LinkedIn

Projects

Fern Bio

AI patent assistant built from the ground up. I worked on data collection, data cleaning, dataset creation, model fine-tuning, and agent creation.

AI agents, fine-tuning, patent workflows

OmnimatteZeroEfficient

Efficient implementation work around OmnimatteZero, focused on fast training-free video matting with pretrained video diffusion models.

computer vision, video diffusion, optimization

RF-DETR Mac

Mac-focused work on RF-DETR for real-time object detection and segmentation workflows.

object detection, segmentation, Apple Silicon

RLM

Regularised linear model experiments and ML fundamentals work, part of my older but still useful learning and implementation base.

machine learning, regularisation, modelling

alchemy

Agent harness experiments for composing multi-step AI workflows.

agents, orchestration, tools

sRAG

Retrieval-augmented generation experiments for semantic search and question answering.

RAG, retrieval, semantic search

Robotics and Hardware

Hardware is a serious side of my work. I have built from scratch harware projects rangings from robot arms, self-balancing bots to autonomous warehouse management swarm robots. I am currently exploring the intersection of robotics and AI, with efficient edge inference as focus.

Publication

Efficient on Device Object Detection for Prosthetic Arm

Peer-reviewed paper on efficient on-device object detection for assistive prosthetic systems.

What I Work On

Data collection, cleaning, and dataset creation for domain-specific AI systems.
LLM fine-tuning, lightweight model adaptation, and practical inference workflows.
Retrieval systems, agent harnesses, and multi-step AI product workflows.
Computer vision systems for video matting, detection, segmentation, and tracking.
Robotics and hardware systems, especially where perception and AI meet the physical world.

Selected Tools

Python, PyTorch, Transformers, RAG, scraping, dataset pipelines, model fine-tuning, computer vision, agents, Apple Silicon inference, robotics, embedded AI.

Contact

I am most active on X, which is also the best place to contact me. You can also find my code on GitHub.