AI engineer

Uzair Attiq

AI engineer based in Islamabad, Pakistan.

I build AI systems: LLM agents, RAG pipelines, and neural networks.
AI degree, plus Stanford and DeepLearning.AI Machine Learning and Deep Learning specializations.
MERN roots, AI focus.

0+Projects Completed
0Certifications
0+Years Building Software
Resume
Portrait of Uzair Attiq

Credentials & Specializations

Certifications

Selected work

Projects

Post Studio

Post Studio

Autonomous LinkedIn AI Publishing Agent (Human-in-the-Loop)

01

An end-to-end agentic publishing pipeline built with LangGraph and Groq LLMs that drafts content, routes it to Slack for human review, and publishes approved posts to LinkedIn. Modular tools run as independent MCP servers over stdio, paired with automated OAuth 2.0 token refreshes and dual PostgreSQL/SQLite state synchronization. Monitored with Langfuse for observability and deployed on Render with an interactive React workflow dashboard.

PythonLangGraphMCPGroq APIFastAPIPostgreSQLSQLiteSlack APILinkedIn REST APILangfuseRender
Source Code
Fitness Bot

Fitness Bot

Fitness Knowledge Assistant: Domain-Specific RAG Pipeline

02

A production-ready fitness RAG pipeline featuring custom markdown cleaning, multi-level header chunking, and rich metadata tagging. It combines Qdrant Cloud vector search with cross-encoder reranking to deliver pinpoint domain context for high-accuracy answers. Powered by Groq and LangChain LCEL via an async FastAPI backend and a live React frontend deployed on AWS.

PythonFastAPIQdrant CloudLangChain (LCEL)Sentence-TransformersCross-EncodersGroqDocker
Live Demo Source Code
Anime Tracker

Anime Tracker

Retro-Styled Watchlist Manager

03

A retro-themed anime tracking web app that organizes watched, completed, and in-progress titles with fluid Framer Motion animations. Features real-time cloud data synchronization powered by Supabase alongside an offline-first local fallback so the watchlist never loses state.

ReactJavaScriptViteTailwind CSSFramer MotionSupabaseCapacitor
Source Code
MOBILE WORKOUT LOGGER

Workout Logger

Mobile-First Fitness Tracker

04

A mobile-first workout logger tailored for rapid exercise logging with intuitive gesture-based digit inputs and historical volume tracking. Engineered with React and TypeScript, running completely offline and packaged directly into an Android APK with Capacitor.

ReactTypeScriptViteTailwind CSSCapacitorAndroid (Gradle)
Source Code

What I work with

Technical Skills

AI & LLM

  • LangChain (LCEL)
  • LangGraph
  • MCP
  • RAG
  • Agentic AI / Tool Calling
  • Vector Databases & Embeddings
  • Cross-Encoder Reranking
  • LLM Evaluation
  • Langfuse
  • OpenAI API
  • Groq API
  • Gemini Vision API

Machine Learning

  • TensorFlow
  • Keras
  • Scikit-Learn
  • NumPy
  • Pandas
  • Neural Networks (CNN, RNN, LSTM, Transformers)

Backend & Frontend

  • Python
  • TypeScript
  • JavaScript
  • FastAPI
  • Node.js
  • Express.js
  • React.js
  • Tailwind CSS
  • REST APIs
  • PostgreSQL
  • MongoDB
  • SQLite

Cloud & Tools

  • AWS (EC2, S3, CloudFront)
  • Render
  • Vercel
  • Docker
  • Nginx
  • Qdrant
  • Supabase
  • OAuth 2.0
  • Git
  • GitHub

Academic Background

Education

Bachelor of Science in Artificial Intelligence

Bahria University, Islamabad, Pakistan

CGPA: 3.5 / 4.0

Relevant Coursework

Artificial IntelligenceMachine LearningDBMSData Communication & NetworkingSoftware Engineering

One last thing

Let's build something that learns.

“The best systems aren’t finished at launch. They get better every time someone uses them.”

LinkedInGitHubEmail