initializing_portfolio.exe
0%
// Full-Stack Developer & AI/ML Engineer

Muhammad
Shaheer
Zaman Shah

>

Building intelligent systems at the intersection of AI, agentic pipelines, and full-stack engineering. CS graduate from FAST-NUCES Islamabad.

Muhammad Shaheer Zaman Shah
01 / Skills

Technical Arsenal

Spanning AI/ML research to production full-stack systems โ€” I work across the entire stack with a focus on agentic architectures and intelligent pipelines.

๐Ÿง 
AI / ML / Computer Vision
TensorFlow LangChain LangGraph MediaPipe ChromaDB HuggingFace Groq scikit-learn FAISS Pandas/NumPy
โš›๏ธ
Frontend
React.js TypeScript Tailwind CSS Vite HTML5/CSS3 Axios WebSocket
โš™๏ธ
Backend
FastAPI Node.js Express.js REST APIs MongoDB MySQL Python
๐Ÿ’ป
Languages
Python JavaScript (ES6+) TypeScript C/C++ Java Assembly
๐Ÿ› ๏ธ
DevOps & Tools
Git/GitHub Docker Postman WordPress MATLAB Cisco EIGRP/OSPF
๐Ÿค–
Agentic Systems
Multi-Agent RAG LangGraph FSM Vector DBs BM25 Retrieval LLM Pipelines RRF Fusion
02 / Projects

Featured Work

video repo watch demo
01 โ€” Enterprise AI
NexusIQ โ€” Enterprise Knowledge Agent
Production-grade multi-source RAG platform ingesting Confluence, Notion, Slack, Jira, Google Drive, SharePoint, and PDFs. 10-node LangGraph pipeline with pgvector HNSW + BM25 RRF, cross-encoder reranking, and hallucination guard for quick, grounded answers.
10-node pipeline 8 data sources
LangGraph PostgreSQL/pgvector FastAPI Next.js 14 Redis Groq
deployed repo click me
02 โ€” Enterprise AI
EMKOS โ€” Multi-Agent Knowledge OS
Enterprise knowledge platform with LangGraph multi-agent orchestration, hybrid BM25 + Pinecone RAG, JWT/RBAC across 4 roles, and SSE-streamed answers. React 19 dashboard with document upload, audit logs, and single-container Docker deploy on Render.
5 agent routes Hybrid RAG + RRF
LangGraph FastAPI React 19 Pinecone Groq PostgreSQL
video repo watch demo
03 โ€” Agentic AI
AgentOps โ€” Eval & Observability Platform
Self-hostable platform for LangGraph/LangChain agents with trace ingestion, nested waterfall views, LLM-as-judge eval graders, regression detection, A/B testing, metric-based alerting, and a GitHub Action CI/CD merge gate.
LLM-as-judge evals CI/CD merge gates
FastAPI LangGraph SDK Next.js PostgreSQL Redis Celery
video repo watch demo
04 โ€” Agentic AI
Visual Novel Video Generator
Fully autonomous pipeline converting natural-language prompts into 50โ€“70s cinematic videos via LangGraph DAG with 4 specialized agents (Story, Audio, Video, Edit). Natural-language edits with multi-level undo and 70+ passing tests.
70+ tests 10 edit intents
LangGraph Groq Llama 3.3 Pollinations.ai React/TypeScript MoviePy Docker
repo
05 โ€” Agentic AI
PodClips โ€” AI Podcast Shorts Generator
Full-stack pipeline extracting viral vertical clips from long-form video/audio. WhisperX transcription, DSP prosody analysis, YAMNet event detection, Ollama LLM clip selection with JSON repair, and GPU-accelerated FFmpeg rendering with karaoke captions.
5-stage pipeline GPU auto-detect
FastAPI WhisperX Ollama FFmpeg MediaPipe Librosa
repo
06 โ€” Agentic AI
NexusSDR โ€” AI Sales Development Rep
Production-grade B2B sales automation agent with LangGraph state-machine workflow, hybrid ChromaDB/pgvector RAG for sales context, 4-stage personalized cold email sequences, and human-in-the-loop approval via graph interrupts before CRM sync.
HITL graph interrupts Hybrid RAG + RRF
LangGraph FastAPI Next.js ChromaDB Groq SQLAlchemy
deployed repo click me
07 โ€” Agentic AI
Lead Generation Agent
8-agent LangGraph pipeline that turns a plain-English client description into verified prospects with personalized outreach. Real lead discovery via Tavily + Apify, domain/email/LinkedIn verification, fit scoring, and live SSE dashboard with CSV export.
8 specialized agents Real lead verification
LangGraph Groq Tavily Apify FastAPI Docker
repo
08 โ€” FYP
VOICE: Pakistan Sign Language Recognition
Real-time PSL recognition system using MediaPipe landmark extraction and a hybrid TCN+Transformer model achieving 98.25% test accuracy across 188-D feature vectors.
98.25% accuracy 100% top-3 accuracy
TensorFlow MediaPipe FastAPI React/Vite TCN+Transformer
video repo watch demo
09 โ€” Agentic AI
Multi-Agent Research & Coding System
Production pipeline converting natural-language tasks into verified Python code via 9 single-responsibility agents with sandboxed execution and hybrid RAG retrieval.
9 specialized agents RRF + Cross-encoder
LangChain FAISS BM25 HuggingFace FastAPI Streamlit
repo
10 โ€” Agentic AI
Agentic Research Engine
Long-form research system with LangGraph state machine, quality-gated retrieval, weighted score fusion (0.7 dense / 0.3 sparse), and consulting-style sectioned report generation.
Hybrid retrieval Auto-retry loop
LangGraph ChromaDB Groq HuggingFace Streamlit
repo
11 โ€” RAG
AI Customer Support Agent
Production-grade RAG system for enterprise customer support with query rewriting, 8-category classification, multi-query hybrid retrieval (dense + keyword RRF), LLM reranking, context compression, and source-grounded citations.
8 query categories Hybrid RRF search
LangChain ChromaDB OpenAI Streamlit tiktoken
video repo watch demo
12 โ€” Full-Stack
Lyrics Finder & Music Critic App
Full-stack music discovery platform with Spotify OAuth, audio feature analysis, AI-generated critiques, per-user ratings in MongoDB, and a complete CI/CD pipeline.
Spotify OAuth Live deployed
React Node.js/Express MongoDB Spotify API
13 โ€” Full-Stack
Trip Management System
Full-stack trip planner with real-time weather forecasts, GPS-based location tracking, and a RESTful FastAPI backend on a normalized MySQL schema with one-click route lookup.
Real-time weather GPS tracking
React.js FastAPI MySQL Weather API GPS API
03 / About

The Person Behind the Code

I'm a Computer Science graduate from FAST-NUCES Islamabad, specializing in AI/ML systems and full-stack development. I build things that actually work โ€” from real-time sign language recognition to fully autonomous video generation pipelines.

My focus is on agentic AI: systems that can plan, retrieve, reason, and act with minimal human intervention. I use LangGraph, LangChain, and Groq to architect multi-agent pipelines that ship to production, not just demos.

Beyond AI, I'm a full-stack engineer comfortable across the entire stack โ€” FastAPI backends, React frontends, MongoDB/MySQL databases, and WebSocket real-time systems.

When not coding: event organiser for university societies, competitive programmer, reader, traveller, and occasional sketcher.

shaheer.profile.ts
class ShaheerZamanShah {
 
// Building AI that thinks & acts
 
name = 'Muhammad Shaheer Zaman Shah'
location = 'Islamabad, Pakistan ๐Ÿ‡ต๐Ÿ‡ฐ'
university = 'FAST-NUCES'
email = 'shaheerzaman023@gmail.com'
 
focusAreas() {
return [
'Agentic AI Systems',
'Multi-Agent RAG Pipelines',
'Computer Vision',
'Full-Stack Development'
]
}
 
languages() {
return {
fluent: ['English', 'Urdu'],
speaks: ['Pashto', 'Punjabi']
}
}
 
interests() {
return [
'Open-source',
'Competitive Programming',
'Reading & Travelling',
'Sketching'
]
}
}
04 / Experience

Work History

JUNE 2026 โ€” PRESENT
Agentic AI Engineer
Upwork ยท Remote
  • Provide agentic AI services to international clients on Upwork โ€” building multi-agent pipelines, RAG systems, and autonomous AI workflows tailored to client needs.
  • Design and deliver production-ready LangGraph/LangChain solutions including lead generation agents, knowledge bases, and custom automation tools.
JULY 2025 โ€” SEPTEMBER 2025
Web Development Intern
Digital Empowerment Network ยท Islamabad
  • Developed and maintained responsive web pages using HTML, CSS, JavaScript, and React.js, improving UI consistency across the platform.
  • Collaborated with a team of developers to deliver features on schedule, participating in code reviews and Agile sprints.
  • Gained hands-on experience with REST API integration and front-end performance optimisation.
2023 โ€” PRESENT
Freelance Web Developer
Fiverr ยท Remote
  • Deliver custom web development projects for international clients โ€” landing pages, portfolios, and small-business websites.
  • Manage the full project lifecycle: requirements gathering, design, development, testing, and client handoff.
2022 โ€” June 2026
Bachelor of Science in Computer Science
FAST-NUCES, Islamabad
2020 โ€” 2022
Intermediate Certificate (HSSC)
FBISE
05 / Contact

Let's Build Something

Open to full-time roles, internships, and freelance projects in AI/ML engineering and full-stack development. Let's connect.