Muhammad Adil Usmani

AI/ML Developer

Building intelligent systems that solve real-world problems through data-driven insights and innovative algorithms. Specializing in machine learning, natural language processing, and production-ready AI applications.

Scroll to explore

Technical Expertise

Mastering cutting-edge technologies to build intelligent solutions

Languages

🐍Python

95%

Primary language for ML/AI development

🗄️SQL

85%

Database queries and data analysis

Backend & APIs

⚡FastAPI

88%

High-performance web APIs

🔗LangChain

85%

LLM application framework

🕸️LangGraph

82%

Graph-based LLM workflows

🔍LangSmith

80%

LLM observability and debugging

🌐REST APIs

87%

RESTful API design and development

⚙️Async Programming

83%

Asynchronous programming patterns

AI/ML & Advanced Techniques

🔄RAG Pipelines

88%

Retrieval-Augmented Generation systems

🤖LLM Applications

87%

Building with large language models

📊Embeddings

85%

Vector embeddings and representations

🔍Retrieval Systems

86%

Information retrieval and search

💡Prompt Engineering

85%

Effective prompt design strategies

🧠Deep Learning

82%

Neural networks and deep architectures

📈Time-Series Forecasting

80%

Temporal data prediction

🔀Hybrid Retrieval

84%

Combined vector and graph retrieval

📊Evaluation Benchmarks

83%

LLM evaluation and benchmarking

Databases

💾SQLite

85%

Lightweight SQL database

🔍ChromaDB

84%

Vector database for embeddings

📊Neo4j

82%

Graph database for relationships

Libraries & Frameworks

🐼Pandas

92%

Data manipulation and analysis

🔢NumPy

90%

Numerical computing

📈Matplotlib

86%

Data visualization

🔥PyTorch

85%

Deep learning framework

⚡TensorFlow

83%

ML and deep learning platform

🧠Scikit-Learn

88%

Machine learning algorithms

Tools & Infrastructure

🐙GitHub

90%

Version control and collaboration

🐳Docker

85%

Container orchestration

⚙️GitHub Actions

84%

CI/CD automation

🚀Render

83%

Cloud deployment platform

▲Vercel

87%

Frontend deployment platform

☁️Azure

78%

Microsoft cloud services

🌐GCP

75%

Google Cloud Platform

✨GitHub Copilot

82%

AI code assistant

🎯Gemini CLI

80%

AI command line tools

📚Deepwiki

80%

Project documentation platform

🔄Airflow

78%

Workflow orchestration

Featured Projects

Showcasing production-ready applications and innovative solutions

Featured Projects

Featured

AeroSphere — Air Quality Forecasting System

Built LSTM model forecasting 72-hour PM2.5 across 45 cities achieving 85%+ accuracy, 30% above baseline. Designed Airflow pipeline processing 1.2M+ records with GPT-based natural language summaries for public air quality reports. Contributed to model training, batch scheduling, and end-to-end Azure cloud deployment.

NASA TEMPO

LSTM

Airflow

Azure

GPT

85%+

accuracy

cities

Featured

Hybrid Graph RAG vs Vector RAG

Comparative analysis of hybrid retrieval from ChromaDB and Neo4j over SEC 10-K filings. Applied late fusion to merge vector and graph contexts for improved structured financial reasoning. Built 53-question benchmark evaluating accuracy, comprehensiveness, diversity, empowerment, and directness metrics.

ChromaDB

Neo4j

Azure OpenAI

LLM-as-Judge

Hybrid RAG

53 Q&As

benchmark

Outperforms

accuracy

Featured

Corrective RAG (CRAG)

Adaptive self-correcting RAG pipeline with threshold-based routing. Routes CORRECT answers (≥0.7) directly, triggers Tavily web search for INCORRECT (<0.3), and runs parallel operations for AMBIGUOUS responses. Optimized for 3-6s latency on direct retrievals and 5-8s with web-augmented answers over geopolitical documents.

LangGraph

FAISS

Tavily Search

GPT-4o-mini

RAG

3-8s

latency

95%+

accuracy

Other Projects

CricTalk

Production-ready cricket knowledge chatbot API with 6 specialized expertise domains. Features enterprise-grade security, API key authentication, rate limiting, and Redis-backed caching.

FastAPI

Google Gemini AI

Redis

LangChain

Security

95%

accuracy

99.9%

uptime

Article Summarization

NLP model using transformer architecture to generate concise summaries of long-form articles with 85% accuracy using T5 small transformer model.

NLP

Transformers

Python

85%

accuracy

10K+

articles

Customer Churn Predictor

ML model predicting customer churn with 92% accuracy using ensemble methods and feature engineering.

Random Forest

Python

92%

accuracy

Predictions

predictions

Stock Price Prediction

LSTM neural networks predicting stock prices using historical data and technical indicators.

Deep Learning

LSTM

TensorFlow

78%

accuracy

Predictions

predictions

Get In Touch

Ready to collaborate on innovative AI/ML projects? Let's build something amazing together.

Or connect directly:

GitHub