Available for new projects

Aaabad Touk

@aaabadcode

AI engineer building intelligent systems — from LLMs and computer vision to production ML pipelines. Turning research into products.

View ProjectsGet in Touch
scroll
// about

Building AI That Ships

Not just research — production systems that scale.

Aaabad Touk
terminal
$ whoami
> aaabad_touk

$ cat skills.txt
> pytorch, langchain, nextjs,
> kubernetes, python, typescript

$ echo $STATUS
> building the future

I'm Aaabad Touk, an AI engineer based in San Francisco. I specialize in building end-to-end machine learning systems — from training custom models to deploying them at scale.

My work spans large language models, computer vision, and MLOps. I believe the best AI isn't the most complex — it's the one that reliably ships value to users.

When I'm not training models or optimizing inference pipelines, you'll find me contributing to open-source ML tools and writing about practical AI engineering.

30+

Projects Shipped

5+

Years Experience

100+

Models Trained

∞

Papers Read

// projects

What I've Built

A selection of AI-powered products and tools.

NeuroChat

An enterprise LLM chatbot platform with multi-model orchestration, RAG pipelines, and real-time streaming. Handles 50k+ concurrent conversations.

ReactPythonLangChainRedisPostgreSQL

VisionLab

Computer vision toolkit for real-time object detection, image segmentation, and visual search. Deployed across 200+ edge devices.

PyTorchOpenCVFastAPIDockerONNX

CodeSensei

AI-powered code reviewer that catches bugs, suggests optimizations, and enforces style guides. Integrated with GitHub Actions.

TypeScriptGPT-4GitHub APINext.jsVercel

FlowML

ML pipeline orchestrator for training, evaluation, and deployment. Supports distributed training across multi-GPU clusters.

PythonKubernetesApache AirflowGogRPC

PixelForge

AI image generation platform with fine-tuning, inpainting, and style transfer. Serves 10k+ generations daily.

Stable DiffusionReactNode.jsAWSS3
// skills

Tech Stack

The tools and technologies I work with daily.

AI / Machine Learning

PyTorch95%
LangChain90%
Transformers88%
Computer Vision85%
MLOps82%

Web Development

TypeScript93%
React / Next.js92%
Node.js88%
Python95%
Go75%

Infrastructure

Docker90%
Kubernetes85%
AWS88%
CI/CD87%
PostgreSQL83%
// experience

Where I've Worked

From open-source contributions to scaling AI at top companies.

2023 — Present

DeepMind

Senior AI Research Engineer

Leading research on multi-modal foundation models. Built internal tooling for large-scale distributed training across TPU pods.

JAXTPUTransformersPython
2021 — 2023

OpenAI

ML Platform Engineer

Designed and scaled inference infrastructure for GPT models. Reduced latency by 40% through model optimization and caching strategies.

PythonKubernetesCUDARedis
2020 — 2021

Scale AI

Full-Stack ML Engineer

Built data labeling pipelines and quality assurance systems for computer vision datasets. Shipped annotation tools used by 5k+ labelers.

ReactPythonFastAPIPostgreSQL
2019 — 2020

Hugging Face

Open Source Engineer (Intern)

Contributed to the Transformers library. Implemented model architectures and wrote documentation for community adoption.

PyTorchTransformersTypeScriptDocker
// blog

Writing & Thinking

Thoughts on AI engineering, ML systems, and building products.

LLM

Why RAG is Eating the AI World

Retrieval-Augmented Generation is becoming the default architecture for enterprise AI. Here's why it works and how to build one.

Mar 20268 min read
ML

Fine-Tuning vs. Prompting: A Practical Guide

When should you fine-tune a model versus engineering better prompts? I break down cost, performance, and maintenance trade-offs.

Jan 202612 min read
MLOps

Building ML Pipelines That Don't Break at 3AM

Lessons from running production ML systems. Monitoring, rollback strategies, and why you need feature stores.

Nov 20256 min read
CV

The State of Computer Vision in 2025

From YOLO to vision transformers — a survey of what's working in production computer vision systems today.

Sep 202510 min read
// contact

Let's Build Something

Have a project in mind? I'm always open to discussing new opportunities.

hello@aaabadcode.dev
GitHubTwitter