Hi, my name is
Sanan Moinuddin
Computer Science graduate student with strong foundation in machine learning and software engineering. Experience developing ML systems using PyTorch, TensorFlow, and scikit-learn. Seeking Summer 2026 SWE/ML internship.
01. About Me
Skills & Expertise
Computer Science graduate student with strong foundation in machine learning and software engineering. Experience developing ML systems using PyTorch, TensorFlow, and scikit-learn for forecasting and recommendation systems. Proficient in Python, FastAPI, and deploying models on cloud platforms. Seeking Summer 2026 SWE/ML internship.
Languages
Development
Tools
Concepts
02. Experience
Where I've Worked
Software Engineer
PepsiCo
- Developed Python automation framework using Selenium WebDriver with Page Object Model design pattern, reducing manual testing effort by 90% and saving 40+ hours weekly across the QA team.
- Designed Azure DevOps CI/CD pipeline optimization by parallelizing test execution, reducing test execution time from 4 hours to 1.5 hours and enabling faster release cycles.
- Built RESTful microservices using FastAPI with PostgreSQL and Redis caching for test data management, improving test throughput by 65% across distributed environments.
- Led comprehensive regression testing across 500+ scenarios using data-driven testing methodology, achieving 99.5% automation coverage with minimal to zero critical production defects.
03. Projects
Featured Work
Project 01
Predictive Infrastructure Scaling System
ML-powered infrastructure scaling system using ensemble models for probabilistic load forecasting.
- Built multi-model ML ensemble combining PyTorch Transformer, XGBoost/LightGBM/CatBoost, and Prophet with quantile regression for probabilistic load forecasting 15 minutes to 7 days ahead.
- Engineered 100+ feature pipeline with cyclical time encodings, rolling statistics, Fourier components for seasonality detection, and business event modeling with prediction calibration for uncertainty quantification.
- Implemented real-time streaming using Kafka for metrics ingestion from Prometheus/Kubernetes, with async processing and dead letter queue for fault tolerance.
- Designed multi-objective decision engine with risk assessment evaluating cost, stability, and spot interruption probability across AWS/GCP/Azure to optimize instance allocation.
Project 02
Multimodal Short Video Recommender System
Content-aware video recommendation system using multimodal neural networks.
- Designed a two-tower neural network that combines ResNet-50, Wav2Vec2 and BERT encoders with gated multimodal fusion for content-aware modality weighting across video, audio, and text.
- Achieved efficient transfer learning by freezing pretrained backbones (ResNet-50, Wav2Vec2, BERT) and training only fusion and projection layers, reducing memory usage and training time.
- Implemented vector similarity search using PostgreSQL with pgvector cosine distance operators and async database operations for scalable video retrieval.
- Built production-ready FastAPI service with batch indexing, top-K retrieval, horizontal scaling support, and cross-platform GPU acceleration (CUDA/MPS/CPU), containerized with Docker.
Project 03
Movie Social Platform
Full-stack social media platform integrating TMDB API for movie discovery, reviews, and community engagement.
- Built a RESTful API with 40+ endpoints handling authentication, user management, posts, comments, movies, and real-time notifications using Express.js and MongoDB.
- Implemented secure JWT-based authentication with Role-Based Access Control (RBAC) supporting user, moderator, and admin roles, plus password reset functionality via email.
- Designed threaded comment system supporting nested replies with full CRUD operations, and integrated TMDB API for movie search, trending content, and genre-based discovery with local database caching.
- Built notification system with MongoDB TTL indexes for automatic expiration and deployed using Vercel (frontend) and MongoDB Atlas with cloud hosting for backend.
Project 04
Kambaz Learning Management System
Full-stack LMS web application enabling course management, user enrollment, and assignment tracking for students and faculty.
- Built RESTful API with Express.js featuring 30+ endpoints for user authentication, courses, modules, assignments, and enrollments using DAO pattern.
- Implemented session-based authentication with MongoDB session store, secure cookies, and Role-Based Access Control (RBAC) supporting Student, Faculty, and Admin roles.
- Designed responsive frontend using Next.js 14 App Router, TypeScript, and Redux Toolkit for centralized state management with dynamic course pages and nested routing.
- Configured production deployment with CORS policies, proxy trust, and cross-origin session handling for Vercel/Render hosting.
Project 05
Calendar Application
Full-featured calendar application built with Java, implementing MVC architecture and industry-standard software engineering practices.
- Architected MVC design with strict separation of concerns, applying SOLID principles and design patterns (Command Pattern for text parsing, Strategy Pattern for export formats) enabling extensible feature development.
- Implemented multi-interface support with Java Swing GUI, interactive CLI, and headless scripting mode, featuring multi-calendar management with IANA timezone support and automatic conversions.
- Built flexible event scheduling system supporting single and recurring events with weekday selection, occurrence limits, and end dates, plus three-level series editing (single, from-this-forward, entire series).
- Developed cross-calendar event copying with date range support and export functionality to Google Calendar-compatible CSV and iCalendar (.ics) formats with comprehensive JUnit test coverage.
04. Education
Academic Background
Northeastern University
Master of Science in Computer Science
Osmania University
Bachelor of Engineering in Computer Science & Engineering
05. Contact
Get In Touch
I'm currently looking for new opportunities. Whether you have a question or just want to say hi, I'll get back to you as soon as possible!