Marefat Abbas Sayyed

About Marefat Abbas Sayyed

I build production-grade machine learning systems that operate at scale.Currently working as an Applied Data Scientist, I design and own end-to-end ML pipelines powering logistics operations processing over 1M parcels daily. My work spans from multi-source data engineering and feature design to model development, AWS-native deployment, monitoring, and runtime optimization.I specialize in building reliable predictive systems rather than experimental models.Over the past 3+ years, I have:• Designed large-scale ETA and delivery date prediction systems achieving ~92% accuracy• Built SageMaker-based batch inference pipelines serving high-volume operations• Engineered structured features across Athena and Redshift datasets• Implemented drift-aware monitoring and fallback modeling for unstable scenarios• Containerized ML services using ECS Fargate with CI/CD integration• Optimized high-volume Pandas workloads using multiprocessing and memory tuningMy modeling expertise includes regression systems, XGBoost, ensemble methods, clustering, and outlier stabilization. My engineering focus is on scalability, maintainability, and operational reliability.What differentiates me is my systems engineering foundation in Electronics, Robotics, Embedded Systems, and Digital Signal Processing. I think in architectures, constraints, and system behavior under failure conditions.I do not build isolated notebooks.I build scalable ML systems designed for real-world uncertainty.Technical Focus Areas:Machine Learning | Production ML | AWS MLOps | SageMaker | Batch Inference | Feature Engineering | Large-Scale Data Processing | Predictive Logistics SystemsI am interested in high-impact roles where I can contribute to scalable AI platforms, architect cloud-native ML systems, and evolve toward senior and principal-level ML engineering responsibilities.

Experience

1. System Engineer Applied Data Scientist (Production Ml Systems)

   Tata Consultancy Services

   Aug 2022 — Present · Pune, IN

   Designing and owning large-scale predictive systems powering nationwide logistics operations processing ~1M parcels daily.Working within a lean ML team while independently managing full ML lifecycle from data engineering and feature design to model deployment, monitoring, and runtime optimization on AWS.Key Contributions• Architected end-to-end ML pipelines for ETA and delivery date prediction• Built regression and XGBoost-based models achieving ~92% prediction accuracy• Enabled 2-hour delivery windows for ~40% shipments• Designed batch inference pipelines using AWS SageMaker• Engineered multi-source structured features using Athena and Redshift• Implemented percentile-based outlier stabilization for prediction reliability• Built fallback-aware modeling logic for unstable operational scenarios• Developed ML-driven pre-delivery notification engine deployed via ECS Fargate• Integrated CloudWatch monitoring, drift checks, and runtime alerting• Optimized high-volume Pandas workloads using multiprocessing and memory tuning• Collaborated directly with Product and Operations teams for real-world deploymentMajor Systems Delivered• Delivery ETA Prediction System• Delivery Date Prediction Pipeline (~1M records/day)• Delivery Timeslot Refinement Engine• Intelligent Delivery Fallback Recommendation System• ML-based Push Notification Engine

Education

• J.T.S.S.P Mandals Shri Shiv Chhatrapati College of Arts, Science & Commerce, Bodkenagar, Junnar Dist. Pune 410 502

  Computer Science

  2018

• Shankarrao Butte Patil English Medium School

  SSC

  2016

• Government College of Engineering And Research,Pune

  Bachelor of Engineering - BE, Electronics and Telecommunication Engineering

  2018 — 2022