Muhammad Usman
AI/ML Engineer + , and Web Engineer
Artificial Intelligence & Machine Learning
Applied machine learning systems, deep learning pipelines, generative AI solutions, and automated analytics.
Automated Grading & Learning Analytics
Intelligent automated grading system using rule-based logic and data analytics to evaluate student performance, reduce manual grading effort, and provide learning insights.
Intelligent Document Summarizer API
AI-powered REST API that summarizes PDF/DOCX/TXT documents, extracts keywords, and answers natural language questions using Groq LLaMA 3.3 70B with offline fallback.
Predicting Students Career Aspirations
ML models predicting students' career aspirations from academic and personal data. Random Forest classifier with 80% accuracy, deployed via FastAPI with career recommendations.
Full-Stack & Web Applications
Modern responsive web applications, high-performance banking platforms, e-commerce solutions, and interactive tools.
CoreBank — NextGen Enterprise Digital Banking Platform
High-performance enterprise digital banking platform featuring SBP-regulated commercial ledgers, real-time PKR liquidity analytics, Raast QR checkout, and PayFast sandbox integration.
Porto's Donuts Bakery — Quetta Artisanal E-Commerce
Modern gourmet bakery e-commerce platform with Quetta express delivery guard, custom donut flavor selectors, persistent Zustand cart management, and instant WhatsApp ordering.
ChronoPulse — Modern Age & Life Insights Calculator
Futuristic cyberpunk glassmorphic web application computing real-time live age down to seconds, 9-planet interplanetary cosmic age, bio-vital heartbeats/sleep statistics, and milestone timelines.
PrimeBrothers E-Commerce Platform
Full-stack e-commerce solution with React frontend, Node.js backend, and PayFast payment integration for a seamless online shopping experience.
Work Experience & Internships
Data Science Intern
✓ Verified TrackDeveloped end-to-end data science solutions including exploratory data analysis, feature engineering, and model building. Built predictive models for disease diagnosis and financial risk forecasting. Deployed models as production APIs using FastAPI and Flask, and created interactive dashboards in Streamlit to communicate insights to non-technical stakeholders.
Certificate of Completion
Machine Learning Intern
✓ Verified TrackBuilt supervised and unsupervised ML models for real-world datasets including sentiment analysis (NLP), anomaly detection in time-series data, and multi-class classification. Worked across the full pipeline — data cleaning, feature selection, model training, evaluation, and API deployment — building a strong foundation in applied machine learning.
Certificate of Completion
Artificial Intelligence Intern
✓ Verified TrackContributed to four team-based deep learning projects — an AI-Powered Interactive Learning Platform, a Super-Resolution Imaging system, an Augmented Reality Try-On experience, and a Smart Parking System with real-time space detection. My focus across all four was building robust data collection and preprocessing pipelines that fed directly into model training workflows.
Certificate of Completion
Hi there
I'm Muhammad Usman — a fresh Machine Learning Engineer and AI Engineer who graduated with a Bachelor of Science in Computer Science from the University of Sindh, Jamshoro in February 2026. During my degree and three back-to-back internships at ITSOLERA PVT LTD, I gained hands-on experience designing, training, and deploying intelligent systems — from classical ML pipelines to deep learning models and LLM-powered applications. Take a look at the tools I work with on my uses page.
Core stack: Python is my primary language across the full ML lifecycle. I build and train models with Scikit-learn, TensorFlow / Keras, and PyTorch, serve them as REST APIs using FastAPI and Flask, and package everything with Docker. For data analysis and visualisation I rely on Pandas, NumPy, Matplotlib, Seaborn, and Plotly. I track experiments with MLflow, manage environments with Conda, and keep all work version-controlled with Git. I am actively expanding my cloud skills on AWS (EC2, S3, Lambda, SageMaker).
What I specialise in: Supervised & unsupervised learning, Natural Language Processing, Computer Vision, Time-Series forecasting, end-to-end MLOps pipelines, and REST API development for ML products. I enjoy bridging the gap between research-quality models and production-ready systems that genuinely help people.
Currently learning: Fine-tuning Large Language Models with HuggingFace PEFT / LoRA, building Retrieval-Augmented Generation (RAG) pipelines with FAISS and Pinecone, and integrating LLM APIs (Gemini, Groq) into real applications. On the infrastructure side I am deepening my knowledge of AWS SageMaker for managed training and deployment, setting up CI/CD pipelines for ML with GitHub Actions and Docker Compose, and exploring real-time Computer Vision with YOLO and OpenCV.