I'm a Machine Learning Engineer who specializes in taking ML models from notebooks to production. Most ML projects never leave Jupyter notebooks—mine actually run in the real world.

• MS in Computer Science @ Purdue University Northwest (GPA: 3.87/4.0) | Graduating May 2025
• Currently working as ML Engineer @ ACCESSIFIERS (Remote, Full-time)
• ICPC Regionalist - Ranked 43/89 teams in USA Mid-Center Regional
• IEEE Xtreme Competitor - Ranked 29/149 teams globally (18th Annual Competition)
• Passionate about MLOps, Production ML Systems, and Full-Stack Development
• Love solving complex problems through code and competitive programming

Production-grade MLOps system predicting NYC taxi demand 15 minutes ahead | 🌐 Live Demo

A complete end-to-end ML system that helps taxi drivers maximize earnings by predicting hyperlocal demand across 250+ NYC geographic regions.

Key Features:

• XGBoost regression model achieving 21.87% MAPE on 10M+ historical taxi trips
• Full MLOps pipeline with DVC for data versioning, MLflow for experiment tracking
• Automated CI/CD using GitHub Actions for zero-downtime deployment
• AWS Infrastructure with Docker, EC2, ECR, and CodeDeploy
• Interactive React Frontend with Leaflet maps for real-time demand visualization
• FastAPI backend serving predictions via REST API with Nginx + SSL

Tech Stack: Python, XGBoost, Optuna, FastAPI, React, Docker, AWS (EC2, ECR, S3, CodeDeploy), DVC, MLflow, GitHub Actions, PostgreSQL, Nginx

Business Impact: Reduces driver empty-mile time by ~20%, decreases customer wait times, increases platform revenue

Imbalanced classification with 284K transactions (0.17% fraud rate)

Built and compared multiple ML models to detect fraudulent transactions with high recall.

Key Features:

• Handled severe class imbalance using SMOTE and NearMiss techniques
• Applied IQR-based outlier removal and feature scaling
• Implemented nested cross-validation for fair model evaluation
• Achieved 81% recall and 94% accuracy using Logistic Regression

Tech Stack: Python, Scikit-learn, Pandas, NumPy, Matplotlib, Seaborn

Models Evaluated: Logistic Regression, K-Nearest Neighbors, Neural Networks

Scalable web application with secure payment integration

Designed and deployed a complete e-commerce platform with authentication, payment processing, and real-time order tracking.

Key Features:

• Reduced page load time by 50% using Node.js REST APIs and React SPA
• Implemented JWT-based authentication with role-based access control
• Integrated secure payment gateway (Stripe) for seamless checkout
• Built real-time order tracking system improving customer transparency
• Applied TDD methodology resulting in 25% reduction in bug reports

Tech Stack: React, Node.js, Express.js, MongoDB, JWT, Stripe API, Redux, Material-UI

• ICPC Regionalist - Ranked 43/89 teams in USA Mid-Center Regional Contest
• IEEE Xtreme - Ranked 29/149 teams globally in 18th Annual Global Competition
• Innovative India Coding Championship - Ranked 1700/90,000 participants
• AWS Cloud Technical Essentials - Coursera (Jan 2024)
• Intermediate Technical Interview Prep - CodePath (Feb-Apr 2025)

Aug 2025 - Present | Remote, Full-time

Leading ML model evaluation and deployment for AI accessibility feedback system helping organizations achieve WCAG compliance.

• Evaluated 25+ LLM models (GPT-4, Claude, Llama) for WCAG criterion inference
• Selected optimal model projected to save $15K+ annually while maintaining 92% accuracy
• Designed comprehensive Website Rating Algorithm aggregating multiple quality metrics
• Built automated evaluation framework reducing manual testing time by 75%

Jan 2025 - May 2025 | On-site, Part-time

Supported 50+ graduate students in Advanced Algorithms and Machine Learning courses.

• Conducted weekly office hours helping resolve 100+ critical bugs in Python, Java, C++
• Mentored students on ML projects involving scikit-learn, TensorFlow, data pipelines
• Improved overall project quality scores by 30% through software engineering best practices

May 2023 - Jul 2023 | Remote, Internship

Developed backend infrastructure for full-stack e-commerce platform serving 10,000+ monthly users.

• Built backend using MVC architecture with Express.js, reducing bug resolution time by 40%
• Implemented 12 RESTful APIs increasing user data retrieval efficiency by 40%
• Applied TDD methodology resulting in 25% reduction in production bugs

I'm always open to interesting conversations and collaboration opportunities!

• Looking for ML Engineer or MLOps Engineer roles
• Open to collaborating on ML systems in production and open-source projects
• Ask me about MLOps, production ML, competitive programming, or full-stack development
• Reach me at: madhajaswanth@gmail.com
• Check out my portfolio: madajaswanth.netlify.app