Open to remote, hybrid, and on-site opportunities

Talha Shahzad

Full Stack Developer | AI / ML Engineer

Building production-grade web, mobile app, AI, and MLOps systems.

Contact Me View Projects LinkedIn Resume

About

A builder, end-to-end.

Full Stack Developer and AI/ML Engineer with 2+ years of hands-on experience delivering production-grade systems across web, mobile apps, and AI. I specialize in applied machine learning and GenAI, from end-to-end MLOps pipelines and multimodal RAG chatbots to computer-vision models, alongside full-stack platforms in Laravel, React.js, and Flutter.

2+ years building production systems end-to-end
Completed 10Pearls Shine Data Science Internship
Production MLOps, time-series forecasting, and explainable AI
10+ web apps shipped (PHP, Laravel, React, WordPress)
Multimodal RAG, LLM, agentic AI, and MCP workflows

Skills

The stack I build with

From machine learning pipelines to frontend — a practical toolkit refined across web, mobile apps, AI, and MLOps.

Languages

PythonJavaScript (ES6+)PHPDartC#C++SQLBashJava

Machine Learning

Scikit-learnXGBoostTensorFlow / KerasLSTMResNet50SHAPPandasNumPy

MLOps

Hopworks Feature StoreModel RegistryFeature PipelinesGitHub Actions CI/CDModel VersioningData Validation

AI / GenAI

RAGLangChainGroq LLaMAFAISSHuggingFaceTensorFlowOCR

Agentic AI / MCP

Model Context ProtocolMCP Servers & ClientsJSON-RPCSamplingRootsClaude Code

Backend / Web

FastAPIFlaskLaravelReact.jsFlutterStreamlit.NETREST APIsOpenAPI

IoT / Embedded

ESP32-S3 CAMArduinoTinyGPS++OpenCVLeaflet.js

Databases

MySQLPostgreSQLSQL ServerFirestoreFAISSSQLite

Cloud & Auth

RenderStreamlit CloudHopworks ServerlessFirebaseGitHub ActionsAWSAzure

Tools & Practices

GitGitHubLinuxPostmanMVCRBACOWASP3NFDSA

Experience

Where I've shipped

Jul 2026 – Sep 2026

Data Science Intern

10Pearls Shine Internship Program — Cohort 9

•Built and deployed Pearls AQI Predictor, an end-to-end MLOps system forecasting Lahore AQI across 24, 48, and 72-hour horizons.
•Engineered 29 temporal, cyclical, lag, rolling, and momentum features from validated hourly weather and air-quality data.
•Compared Ridge Regression, Random Forest, XGBoost, LSTM, and persistence models; selected Ridge with 0.9988 validation R² and 0.9868 test R².
•Productionized Hopworks feature-store and model-registry workflows, SHAP explanations, a Flask API, Streamlit dashboard, and scheduled GitHub Actions pipelines.
•Validated a live 166 AQI prediction against Punjab EPA's official reading of 160, a deviation of about 4%.

Jan 2024 – Present

Full Stack Web, Mobile & AI Developer

Freelance / Self-Employed

•Architected a multimodal RAG chatbot (Flutter + FastAPI + FAISS + Groq LLaMA 3.3 70B + Firebase) with bilingual voice I/O and real-time Firestore history.
•Shipped 10+ production web apps using PHP, Laravel, React.js, and WordPress under MVC, RBAC, and OWASP standards.
•Developed cross-platform Flutter mobile apps with Firebase Auth, Firestore, and Storage for Android and iOS.
•Delivered a retail POS desktop app (C#, WinForms, SQL Server) and a 3NF-normalized Real Estate Rental Management System.

Education

Sep 2023 – 2027 (Expected)

B.S. Information Technology (In Progress)

Ghazi University, Dera Ghazi Khan

Projects

Selected work

Production systems and capstones spanning AI, IoT, and full-stack.

Pearls AQI Predictor

Production MLOps system forecasting Lahore AQI for 24/48/72-hour horizons. It combines 29 engineered features, comparative modeling, Hopworks feature and model registries, recursive forecasting, SHAP explanations, a Flask API, and a Streamlit dashboard with automated GitHub Actions pipelines.

PythonScikit-learnXGBoostTensorFlowHopworksSHAPFlaskStreamlit

View on GitHub Live dashboard

Multimodal RAG Chatbot

Full-stack AI assistant supporting PDF/DOCX/PPTX/XLSX/CSV, OCR images, and MP4 speech-to-text. Bilingual voice I/O (EN/UR), Firebase Auth, FAISS semantic search.

FlutterFastAPIGroq LLaMAFAISSFirebase

View on GitHub

AI Voice Chatbot with RAG

Enterprise RAG chatbot with multi-source ingestion (PDF, DOCX, OCR images, MP4 audio), persistent FAISS index, and ultra-fast Groq LLaMA 4 inference.

PythonStreamlitGroqFAISSTesseractFFmpeg

View on GitHub

POSApp — Retail Point of Sale

Desktop POS for a men's clothing store: product catalog, inventory, billing engine, and a sales reporting dashboard.

C#WinFormsSQL Server.NET

View on GitHub

Real Estate Rental Management

Comprehensive rental platform: tenants, lease lifecycle, property portfolio, and financial reporting on a fully 3NF-normalized schema.

C#SQL ServerWinForms.NET

View on GitHub

Achievements

Certifications & Awards

AI Fluency: Framework & Foundations

Anthropic · Jul 2026

Introduction to Model Context Protocol

Anthropic · Jul 2026

Model Context Protocol: Advanced Topics

Anthropic · Jul 2026

Claude Code 101

Anthropic · Jul 2026

AI For Everyone

DeepLearning.AI · Dec 2024

Networking Basics

Cisco Networking Academy · Dec 2024

Microsoft Azure SQL

Microsoft / Coursera · Dec 2024

Introduction to Cybersecurity

Cisco Networking Academy · Sep 2024

AWS Cloud Fundamentals

Microsoft Azure Fundamentals

Shine Internship Program — Data Science (Cohort 9)

10Pearls · Sep 2026

Contact

Let's build something.

Open to full-time roles, freelance projects, and AI/IoT collaborations.

mtalhashzdkhn53@gmail.com +92 334 657 3003

Dera Ghazi Khan, Pakistan