Senior Software Engineer | Machine Learning Engineer
15 years of production software engineering experience now specializing in Machine Learning and AI. Combining deep technical expertise (Android, .NET, React) with cutting-edge ML skills (TensorFlow, scikit-learn, Deep Learning) to build production-ready AI systems. Led teams, shipped apps with millions of users, and recently built 13+ ML systems in 15 days. Ready to bring both worlds together.
End-to-end ML system predicting customer churn for telecom industry. Analyzes customer behavior, identifies high-risk accounts, and enables proactive retention strategies. Built complete pipeline from data analysis to production deployment with business ROI quantification.
Business Impact: Identifies 83% of churning customers, enabling targeted retention campaigns that could save $300K annually through reduced churn.
Technical Highlights: Engineered 15+ features, implemented risk scoring system (Low/Medium/High), deployed production REST API with comprehensive error handling.
CNN-based image classification system achieving 99.3% accuracy using transfer learning. Compared multiple architectures (MobileNetV2, ResNet50, EfficientNet) and implemented custom training pipeline with data augmentation for improved generalization. Created reusable framework supporting any custom image dataset.
Key Achievement: Achieved state-of-the-art accuracy using transfer learning, reducing training time by 80% while improving performance over from-scratch training.
Technical Implementation: Custom ImageDataGenerator pipeline, architectural benchmarking (MobileNetV2 vs ResNet50 vs EfficientNet), production deployment framework.
Multivariate LSTM model for weather forecasting using correlated features (temperature, humidity, pressure, wind). Implemented sequence-to-sequence architecture for 7-day predictions. Compared LSTM vs traditional ML approaches to demonstrate when deep learning outperforms classical methods.
Technical Highlight: Multivariate approach using 4 correlated features improved accuracy by 20% over univariate baseline. Demonstrated superiority of LSTM for sequential data with long-term dependencies.
Applications: Agriculture planning, energy demand forecasting, event scheduling, disaster preparedness.
Production-grade fraud detection system handling highly imbalanced data (0.17% fraud rate). Implemented SMOTE for class balancing, ensemble methods (XGBoost, LightGBM), and optimized for precision-recall trade-off. Deployed as real-time API with <50ms response time.
Business Value: Detects 89% of fraud cases while maintaining 93% precision, reducing false positives by 40% compared to baseline. Real-time detection prevents fraudulent transactions before completion.
Full-stack ML deployment with automated feature engineering pipeline, RESTful API, and production-grade infrastructure. Implements model versioning, health checks, batch predictions, and comprehensive error handling. Containerized with Docker for easy deployment across environments.
Infrastructure: Containerized deployment with health monitoring, automated feature engineering, model persistence, batch prediction support, and comprehensive API documentation.
Key Projects:
QA background provides strong foundation for ML model validation, A/B testing, and production monitoring
Intensive self-taught transformation from zero ML knowledge to building production-ready AI systems. 60+ hours of hands-on coding covering the entire ML lifecycle from data analysis to deployment.
Python fundamentals, NumPy, Pandas, Data Visualization (Matplotlib, Seaborn), Exploratory Data Analysis, Git version control
Linear & Logistic Regression, Decision Trees, Random Forests, Model evaluation metrics, Train-test splitting
Ensemble Methods (XGBoost, LightGBM), Gradient Boosting, Imbalanced data handling (SMOTE), Feature engineering, Fraud detection (97.8% accuracy)
scikit-learn Pipelines, Feature engineering automation, REST API deployment (Flask), Model persistence (joblib), Production best practices
Neural Networks, Backpropagation, CNNs for image classification, MNIST (98.9% accuracy), CIFAR-10 (78-82% accuracy)
Transfer Learning (MobileNetV2, ResNet50, 99.3% accuracy), LSTMs for time series, Weather forecasting (1.8°C MAE), LSTM vs Traditional ML comparison
Portfolio development, Customer churn prediction (85% accuracy, $300K impact), Resume & cover letter, Job application strategy
GitHub profile polish, Company research & applications (Brain Station 23, Reve Systems), Upwork profile setup, LinkedIn networking strategy
15 years software engineering + Fresh ML expertise = Rare skill set
Most ML engineers struggle with production deployment. Most software engineers don't know ML. I bridge both worlds—building models that actually ship to production, not just notebooks.
99.9% crash-free sessions, 67 req/s APIs, <20ms latency
I know what "production-ready" actually means. Every ML project includes deployment code, error handling, monitoring, and scalability considerations from day one.
Led team of 5 engineers, mentored juniors, shipped to 10K+ users
Not just an individual contributor. Can lead ML teams, mentor developers, and align technical work with business goals. Proven ability to deliver projects on time and within budget.
Zero to production ML in 16 days
Self-taught Android, Flutter, React Native, Spring Boot, Django, and now ML. Proven ability to rapidly master new technologies and apply them to real-world problems.
Every project includes ROI analysis and impact quantification
Not just chasing accuracy numbers. Every ML project includes business impact analysis ($300K churn prevention, 40% false positive reduction). I understand that models exist to solve business problems.
Async communication, documentation-first, self-motivated
Experienced with remote work, comprehensive documentation, and independent execution. 16-day ML journey demonstrates self-discipline and ability to deliver without supervision.
Seeking remote Machine Learning Engineer positions where I can leverage 15 years of production software engineering combined with cutting-edge ML expertise.
Open to: Full-time | Contract | Freelance