Akash Pathak

Machine Learning Engineer with 2+ years of hands-on experience designing, training, and productionizing ML models — from gradient boosting to deep learning and NLP — on large-scale user and transactional data. Skilled at translating experimental notebooks into clean, modular, production-ready Python, and deploying models as low-latency REST APIs using FastAPI and Docker. Experienced building high-throughput feature engineering pipelines, monitoring production models for performance drift, and retraining to sustain accuracy.

Comfortable collaborating closely with backend engineers, data engineers, and product managers in fast-moving, in-office teams.

AI/ML Engineer at RateGain (2025-01 – Present)

• Design, train, and evaluate ML models spanning gradient boosting (XGBoost, Random Forest), deep learning, and NLP frameworks to improve prediction accuracy and downstream decision quality.
• Built large-scale feature engineering pipelines over millions of hotel listing and booking records, combining text normalization, multilingual embeddings, geo-spatial retrieval (BallTree), and FAISS approximate nearest-neighbor search for high-precision entity matching.
• Translated experimental notebook code into clean, modular, production-ready Python services; deployed models as low-latency REST APIs using FastAPI and Docker, processing 2M+ records daily with 35% lower latency through batching, caching, and optimized request handling.
• Containerized and orchestrated production ML services using Docker and Kubernetes (GKE) on GCP Cloud Run, ensuring scalable, reliable serving under production load.
• Set up MLflow-based experiment tracking, model versioning, and performance monitoring to detect degradation and trigger retraining, sustaining model accuracy over time.
• Collaborated cross-functionally with product and engineering stakeholders to scope model requirements, review architecture, and translate business problems into deployable ML solutions.

AI/ML Engineer Trainee at RateGain (2024-07 – 2024-12)

• Performed exploratory data analysis and feature engineering on large-scale transactional booking data using Python and SQL.
• Built and tuned a gradient-boosted (XGBoost) classification model with hyperparameter search to predict booking cancellations.
• Evaluated models using precision-recall, ROC-AUC, and confusion-matrix analysis, and clearly communicated statistical findings to both technical and non-technical stakeholders.

B.Tech in Computer Science and Engineering (AI/ML) – ABES Engineering College, Ghaziabad (2020-09 – 2024-07)