Machine Learning Engineer with 4+ Years of Experience

Kotesh Kumar Yelamati

Machine Learning Engineer Location: NJ Mobile: 401-***-**** Email: ******@*************.*** SUMMARY:

Machine Learning Engineer with 4+ years of expertise in building and deploying end-to-end AI/ML solutions. Skilled in Python, TensorFlow, PyTorch, and NLP frameworks like BERT, GPT, and Llama. Hands-on with cloud platforms (AWS, GCP, Azure), big data tools (Spark, Kafka), and MLOps practices (Docker, MLflow, CI/CD). Proven ability to deliver scalable, production-ready models for automation, prediction, and intelligent decision-making. SKILLS:

Methodology: SDLC, Agile, Waterfall

Programming Languages: Python, R, C++, Java

Machine Learning Frameworks: TensorFlow, PyTorch, Keras, Scikit-learn, Hugging Face Transformers Deep Learning: CNNs, RNNs, LSTM, GANs, VAEs, Diffusion Models Natural Language Processing: BERT, GPT, spaCy, NLTK, Llama, Roberta Data Management: SQL, MongoDB, Cassandra, HBase

Generative AI: Text generation, image synthesis, data augmentation, prompt engineering, fine-tuning large language models Big Data Technologies: Apache Hadoop, Spark, Kafka Cloud Platforms: AWS, Google Cloud, Azure, Vertex AI, AWS SageMaker Data Visualization: Tableau, Power BI

Model Deployment and MLOps: Docker, Kubernetes, CI/CD pipelines, MLflow, TFX EDUCATION:

Master of Science in Data Analytics Apr 2025

Indiana Wesleyan University

EXPERIENCE:

Genworth, NJ Oct 2024 – Current Machine Learning Engineer

• Coordinated cross-functional teams to plan and execute machine learning projects under Agile methodology, conducting sprint planning and retrospectives to boost delivery efficiency by 10%.

• Designed and deployed a predictive risk assessment model using XGBoost and SHAP to forecast long-term care insurance claims, improving early risk detection by 28% and reducing false positives by 15%.

• Built an NLP pipeline with BERT and spaCy to extract insights from unstructured medical documents and underwriting notes, leading to 20% faster claims review by automating feature extraction.

• Integrated a real-time fraud detection system using Isolation Forest and Autoencoder models, identifying anomalous claim patterns and reducing fraudulent payouts by $2M annually.

• Led the migration of model training workflows to AWS SageMaker, implementing CI/CD with AWS Code Pipeline and reducing deployment time from days to hours.

• Developed an Explainable AI (XAI) framework for underwriters using LIME and SHAP, enhancing model transparency and gaining regulatory approval for AI usage in decision-making.

• Collaborated with actuarial and compliance teams to create fair and bias-mitigated ML models using Fairlearn, ensuring demographic fairness across age and gender groups. Infinite Infolab, India April 2020 – July 2023 AI/ML Engineer

• Developed a computer vision-based defect detection system using YOLOv5 and OpenCV for a manufacturing client, achieving 94% accuracy in real-time anomaly detection on the assembly line.

• Built a multilingual chatbot using Rasa and spaCy, integrated with client CRM via REST APIs, reducing support tickets by 35% across English, Hindi, and Tamil users.

• Led an AI-driven customer churn prediction project using LightGBM and survival analysis, improving retention strategies and reducing churn by 22% for an e-commerce client.

• Designed and deployed a recommendation engine using Collaborative Filtering and Matrix Factorization for a retail client, increasing cross-sell revenue by 18%.

• Created time-series forecasting models with ARIMA, Prophet, and LSTM to predict energy demand for a smart-grid solution, achieving MAPE < 7% on unseen data.

• Developed and productionized ML pipelines using MLflow, Airflow, and Docker, ensuring reproducibility and reducing experimentation time by 40%.

• Conducted end-to-end model validation and A/B testing, presenting actionable insights to stakeholders through Power BI dashboards, resulting in data-driven feature deployments in 4 client products. PROJECTS:

Lung Cancer Detection & Classification System using Optimized KNN & Comparative ML Models Published at IEEE ICECONF 2023 — View on IEEE Xplore

• Developed a machine learning pipeline to identify lung cancer cases from patient data using an optimized K-Nearest Neighbors

(KNN) model, improving detection accuracy by benchmarking against Logistic Regression and ensemble classifiers.

• Engineered and preprocessed raw medical datasets by handling missing values, removing outliers, and standardizing features, enabling a reliable and reproducible test-driven model training process.

• Conducted comparative model evaluation (KNN, Decision Tree, Random Forest) using precision, recall, F1-score, and confusion matrices, identifying edge case misclassifications for robust model selection.

• Automated end-to-end model training and validation workflows, documenting all steps from data preparation to hyperparameter tuning, ensuring transparency, reproducibility, and ease of iterative improvement. Automated Chest X-Ray Image Classification using CNN for Pulmonary Disease Detection

• Designed and trained a Convolutional Neural Network (CNN) to classify chest X-ray images such as normal, pneumonia, and tuberculosis, achieving over 92% validation accuracy using augmented image datasets.

• Implemented a Flask-based web interface for doctors to upload X-rays and receive real-time model predictions, later containerized using Docker and deployed on AWS EC2 for scalable access.

• Applied image preprocessing techniques (resizing, histogram equalization, noise removal), along with data augmentation

(rotation, zoom, flipping), to improve model generalization and reduce overfitting. CERTIFICATIONS:

Data Analysis with Python – IBM Developer Skills Network SQL and Relational Databases – IBM Developer Skills Network

Contact this candidate