Skip to content
View MuhammadAamirGulzar's full-sized avatar
8 followers · 4 following

Achievements

Achievement: Quickdraw

Achievements

Achievement: Quickdraw

Highlights

  • Pro

Block or report MuhammadAamirGulzar

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
MuhammadAamirGulzar/README.md

Hi, I'm Aamir 👋

Independent AI Engineer · Production Generative AI & Agentic Systems
I build AI systems that make it to production — not just to a demo.

Portfolio LinkedIn Google Scholar

I'm an AI/ML Engineer who takes LLM and agentic systems from prototype to production, with live solutions across RAG, voice AI, and enterprise automation. I've worked in AI since 2021, and over the past 3+ years I've shipped LLM applications, real-time voice agents, enterprise automation pipelines, and multimodal systems. I turn business problems into working AI systems, choosing the right approach for each, from data processing and feature engineering through model training, optimization, large-scale deployment, and continuous learning. Q1-published researcher in medical AI, bringing research rigor to engineering.

  • 🔭 Currently: Independent AI Engineer, leading development of GovXprt, a production multi-agent RAG system for a confidential US enterprise client, on Vertex AI
  • 💼 Previously: AI Engineer at TalentBridge Oy (Finland, remote), Nov 2025 – Jun 2026 - shipped automated contract generation for client Neste on Vertex AI + Snowflake, plus three production products for TalentBridge itself: SUSAN-GPT, Staffin, and EagleChat
  • 🔬 Research: Joint first-author, Q1 journal (2025) - benchmarking pathology foundation models for MSI prediction in colorectal cancer, plus a first-author Wiley book chapter (2024) on skin-disease detection · Google Scholar
  • 🎓 MS Computer Science, FAST-NUCES · former Lecturer & AI Trainer
  • 🧠 Building: a multi-agent system (agentic orchestration, tool use, evaluation)
  • 🌐 Full portfolio & book a meeting: muhammadaamirgulzar.github.io

🚀 Featured Work

Project What it demonstrates Stack
SUSAN-GPT Voice-enabled RAG career/business assistant for jobseekers in Finland · live in production RAG · Voice AI
Staffin AI resume-matching + multi-agent interview screening platform · live in production Multi-agent · Resume matching
GovXprt AI-assisted regulatory compliance platform for environmental health permitting · live in production RAG · Regulatory Compliance
EagleChat AI tutoring for NCLEX, journalism, medical assistant training, and financial literacy · built on the SUSAN-GPT backend · live in production RAG · Voice AI · EdTech
Contract Automation Pipeline Enterprise doc automation · 98%+ field accuracy · <5 min/trade Vertex AI Pipelines · Gemini 2.5 Pro · Snowflake
Zoning Map Agent Deterministic-first agent that recovers delisted docs · bounded LLM escalation Python · Playwright · Docker
colonmsi-fm Code behind my Q1 pathology-AI publication CONCH · UNI · PyTorch · OpenSlide
Urdu AI Video Creator Full-stack generative video · voice cloning + lip-sync Wav2Lip · XTTS v2 · Next.js · FastAPI

🛠️ Tech I Work With

GenAI & Agents LangChain LangGraph CrewAI LlamaIndex RAG Vertex AI Gemini OpenAI Ollama Hugging Face

Voice & Real-Time AI Pipecat WebRTC Twilio Whisper Deepgram ElevenLabs

ML / DL PyTorch TensorFlow Transformers XGBoost scikit-learn

Backend & Data Python FastAPI Node.js Playwright PostgreSQL MongoDB Redis Snowflake Qdrant ChromaDB

Frontend React Next.js TypeScript Tailwind CSS

Cloud & MLOps GCP Kubeflow Cloud Run AWS Azure Docker

🎯 What I focus on

Agentic AI & RAG      →  LangGraph, multi-agent orchestration, agentic RAG, evaluation (Ragas/DeepEval)
Voice & Multimodal    →  real-time ASR/TTS, WebRTC pipelines, VLMs, generative media
LLM Engineering       →  fine-tuning (LoRA/QLoRA), quantization, on-device/edge inference
AI Research           →  computational pathology, spatial transcriptomics, foundation models

Open to AI engineering and team lead roles, research collaborations, and consulting on GenAI products.

Popular repositories Loading

  1. TextClassification_ML_DL TextClassification_ML_DL Public

    First Text Classification Repo using Machine learning and Deep learning

    Jupyter Notebook

  2. transformer-qa-pipeline transformer-qa-pipeline Public

    Fine-tune and run extractive QA models on SQuAD with reproducible training and inference workflows.

    Python

  3. vib-baseline-wsi-finetuning vib-baseline-wsi-finetuning Public

    Variational Information Bottleneck baseline (Li et al., CVPR 2023) adapted for weakly-supervised WSI classification research

    Python 1

  4. cnn-baseline-wsi-classification cnn-baseline-wsi-classification Public

    ResNet34/ConvNeXt patch-classification baseline for WSI biomarker prediction; comparison baseline in an MSI/MSS research pipeline

    Jupyter Notebook

  5. kat-baseline-zenodo-patches kat-baseline-zenodo-patches Public

    KAT (Kernel Attention Transformer) baseline adapted for a Zenodo pre-patched WSI dataset; part of an MSI/MSS classification research pipeline

    Python

  6. kat-baseline-tcga-wsi kat-baseline-tcga-wsi Public

    KAT (Kernel Attention Transformer) baseline adapted for a TCGA colorectal cancer WSI cohort; part of an MSI/MSS classification research pipeline

    Python