AI engineer · researcher · Karachi, PK

I build AI systems that move from research paper to production — in weeks, not quarters.

Fresh BS in Artificial Intelligence from FAST NUCES with 2+ years shipping real models. Currently an L2 AI Software Engineer collaborating with a Google research lead on multi-agent architectures. Published at IEEE INMIC 2.0, 2nd at TeknoFest 2024.

~90s from idea to prototype LLMs · VLMs · CV 60% → 93% accuracy lift

What I'm working on this season.

Three threads, one question: how far can a small team push AI agents that actually reason?

Sept 2025 — now

Multi-agent architectures under NDA

SleepWellSystem · AI Engineer L2 (contract)

Partnering with a research lead at Google on reasoning and workflow automation for autonomous agents. Designing LLM integrations that hold up outside benchmark conditions.

agents reasoning LLM orchestration
Jan 2024 — now

Climate.io — final year project going to production

FAST NUCES · 2nd at TeknoFest 2024

Water-quality assessment pipeline: VLM extracts context from water imagery, an LLM writes the environmental report. Shipped end-to-end on Hugging Face + FastAPI.

VLM LLM FastAPI climate
Feb 2024 — now

Fine-tuning LLaMA & image-to-text on Intel datasets

Intel Corporation · Student Brand Ambassador

Optimizing model performance and scalability across large Intel-provided datasets. Quietly one of the more educational things I do.

LLaMA fine-tuning i2t

Seven teams. Five time zones. One discipline.

From annotation work at Scale AI to research labs to contract agent engineering — I've seen AI from every angle of the stack.

SleepWellSystem

● Current
AI Software Engineer L2 · Contract

Building multi-agent reasoning systems with a Google research lead. Scalable LLM integrations for real-world workflows.

Onsite · PakistanSept 25 — now

Ahoy

Completed
AI Researcher

CV-based football analysis system with real-time object detection & tracking. Also: an NLP calling bot with speech recognition and dialogue.

RemoteJan — Aug 25

Intel Corporation

● Active
Student Brand Ambassador

Fine-tuning LLaMA and image-to-text models on Intel datasets, optimizing for scalability and downstream deployment.

Remote · CA, USAFeb 24 — now

Axillio

● Active
Junior AI/ML Engineer

Lifted model accuracy 60% → 93% with YOLO, ResNet, Inception. Cut video generation from 2 days to 10 minutes. Built the Node/Express API around it.

Remote · LahoreJan 24 — now

FAST NUCES — Karachi

● Active
Junior AI Researcher

UAV modeling with LSTM, brain-therapy CNN analysis, satellite image dehazing, GAN-based recommenders. Two research codebases on GitHub.

Hybrid · KarachiJun 23 — now

Scale AI

Completed
Senior Data Annotator

Where the discipline for clean training data came from. Senior-tier review and annotation for ML pipelines.

Remote · SFMay — Aug 23

DreamSleepApp

Completed
Software Engineer Intern

Automated video + audio editing workflows with MovieLip and pydub. First taste of end-to-end media pipelines.

Remote · ChicagoAug — Oct 23

+ Open to your team

Remote AI roles

Agent systems, VLM/LLM product work, applied research. Pakistan-based, fully remote, full-time.

AnywhereAvailable now

Selected work, with the parts that hurt left in.

Four featured projects. Each one shipped, each one taught me something I still use.

VLM → LLM contamination: 0.82
★ 2nd place · TeknoFest 2024

Climate.io

Final year project · 2024 — ongoing · IEEE INMIC 2.0 paper

A water-quality assessment pipeline that turns a photo of contaminated water into a readable environmental report. A vision-language model extracts visual evidence, a large language model writes the analysis, and both get packaged into an end-to-end system deployed on Hugging Face and FastAPI.

Also won Ideathon UBIT. Published the comparative analysis between VLMs and LLMs for climate monitoring at IEEE INMIC 2.0.

VLMLLMFastAPIHuggingFaceclimate
> query: "explain GAN collapse"
> reformulated:
  "Describe mode collapse
   in generative adversarial
   networks, with 2 mitigation
   strategies."
> retrieved: 6 docs
> latency: 340ms

ResearchEase

AI research tool · 2024

An auto-prompting layer that rewrites lazy user queries into prompts that actually work. Built on Hugging Face + LangChain + LangSmith + ChromaDB, with a Flask backend and a web UI. Removed the prompt-engineering step for researchers reading across dozens of papers.

LangChainChromaDBLangSmithFlask
COLLA-GAN / ACAE
user × item × context

Recommendation System on GANs

Research project · MovieLens · 2023

Four models compared on implicit feedback — BPR, APR, CollaGAN, ACAE — benchmarked across Precision, Recall, HR, NDCG, MAP, and F1. Full adversarial training loop, evaluation plots auto-exported to PDF. My introduction to reading a recsys paper and rebuilding it from scratch.

CollaGANACAEBPRAPRMovieLens
WhatsApp bot Email
Files Video Certificates
—80% manual work

Automation Suite

Productivity tools · 2023

Five small programs that each killed a recurring task: WhatsApp bot, email automation, file organizer, video editing, certificate generator. Together they cut repetitive work by up to 80% for the teams using them.

Pythonautomationtooling
Happy to walk through these in a call
🔒 Nexus
Multi-agent reasoning system for workflow automation.
agents · LLM
🔒 getspire
Productized AI tooling — internal codebase.
product · AI
🔒 ResearchEase
Auto-prompting layer for literature Q&A (LangChain + ChromaDB + LangSmith).
RAG · research
🔒 Concierge
Conversational assistant with long-context memory.
agents · NLP
🔒 algolark
Algorithmic systems sandbox — applied ML infra.
infra
🔒 DocGen
Document generation pipeline with LLM-assisted layout.
LLM · docs
🔒 Multi_Speaker_Transcript
Speaker diarization + transcription with Whisper pipelines.
speech · Whisper
🔒 RAG-model
Retrieval-augmented generation harness with evaluation.
RAG
🔒 vector-embedding
Embedding service & benchmarks across models.
embeddings
🔒 CV-Matching
Resume-to-JD semantic matcher for hiring pipelines.
NLP · hiring
🔒 Brain-Therapy
CNN analysis for brain-therapy research (FAST lab).
CNN · medical
🔒 PlasticGAN
GAN implementation for plastic-waste imagery.
GAN · climate
🔒 invention
Early-stage prototype — details on request.
exploratory
🔒 Image_Classification
Client image-classification pipeline (Axillio).
CV · production
🔒 Erase-and-Replace
Inpainting pipeline for video editing workflows.
CV · video
🔒 Internship_Work
Consolidated internship deliverables (private).
archive
🔒 Remote_jobs_Ai
Job-market analysis & scraping for AI roles.
data · scraping
🔒 IR
Information-retrieval coursework & experiments.
IR

Published research.

IEEE

A Comparative Analysis Between Visual Language and Large Language Models to Monitor Contaminated Water in the Face of Climate Change

INMIC 2.0 · Paper 2024386402 · Conference 64792X · Accepted, IEEE Xplore
Awaiting Xplore DOI

What I reach for.

Filled dots are the ones I use weekly. The rest I'm fluent in but don't write love letters about.

Languages

  • Python
  • JavaScript
  • C / C++
  • Java
  • AssemblyScript
  • PowerShell

AI / ML

  • PyTorch
  • TensorFlow
  • LangChain
  • Hugging Face
  • scikit-learn · Keras
  • OpenCV · mlflow

Backend

  • FastAPI
  • Flask
  • Django
  • Node · Express
  • JWT · REST

Data & Infra

  • MongoDB
  • MySQL · SQLite · Oracle
  • Docker
  • Git · GitHub
  • Linux · Vercel
  • Pandas · NumPy · Plotly
FAST National University
BS Artificial Intelligence · GPA 3.0 · CV, NLP, ANNs, Parallel Computing
Data Science · ML Specialization · Mathematics for ML
Tech Director, Procom24 · ACM · IEEE

Let's build something that ships.

Looking for a remote AI engineer who has read the papers and also shipped the thing? I'd love to hear what you're working on.

+92 305 2023022 · Karachi, Pakistan · UTC+5