Download My Resume
Souritra Banerjee
Applied Scientist with 5+ years building production ML systems at Amazon. Delivered $500K+ quantified impact through demand forecasting, MLOps automation, and LLM-powered operational intelligence. IIT Madras graduate specializing in end-to-end model development and scalable AI solutions.
CURRENT ROLE
Applied Data Scientist at Amazon
Research Science Ops Tech
Leading ML forecasting systems and LLM-based solutions across global operations. Delivered $450K annual savings through ensemble forecasting models and reduced investigation time by 60% with automated anomaly detection.
  • Improved demand forecasting MAPE from 5.2% to 3.1%
  • Built production MLOps pipelines for 8+ models
  • Reduced model refresh latency by 65%
Key Technical Achievements
ML Forecasting System
Prophet + XGBoost ensemble predicting global demand across 10+ regions with 3.1% MAPE accuracy
LLM Anomaly Detection
LangChain + OpenAI solution processing 50+ daily signals, reducing investigation time by 60%
MLOps Automation
SageMaker pipelines with automated retraining, monitoring, and deployment for production models
Transformer Research
Collaborated on BERT/RoBERTa experimentation achieving 28% recall improvement for NLP tasks
Technical Expertise
ML & Deep Learning
PyTorch, TensorFlow, Scikit-learn, XGBoost, LightGBM, Hugging Face Transformers
LLMs & NLP
Claude, LangChain, RAG pipelines, Prompt Engineering, fine-tuning
MLOps & Engineering
SageMaker, MLflow, Docker, Kubernetes, CI/CD, Model Monitoring
Data & Analytics
SQL (PostgreSQL, Redshift), Python, A/B testing, Time Series Analysis
Cloud Platforms
AWS (SageMaker, Lambda, S3, Redshift, EC2), Git
Impact
Quantified Business Impact
$500K+
Total Impact
Quantified value delivered through ML systems and automation initiatives
$450K
Inventory Savings
Annual cost reduction from improved demand forecasting accuracy
$2M
Revenue Contribution
Annualized impact from A/B testing framework and promotion optimization
65%
Latency Reduction
Model refresh time improvement through MLOps workflow optimization
60%
Time Saved
Investigation time reduction via automated anomaly summarization
8+
Production Models
ML models deployed with automated pipelines and monitoring
Career Journey at Amazon
1
Jun 2024 – Present
Applied Data Scientist - Research Science Ops Tech. Leading ML forecasting systems and LLM solutions. Recognized as Outstanding Performer (2024).
2
Jun 2023 – Jun 2024
Data Scientist - Category Team. Led A/B experimentation framework analyzing 200K+ sellers. Awarded Best Employee (2023).
3
Sep 2022 – Jun 2023
Data Analyst - Logistics Operations. Built delivery optimization model improving on-time rate by 6%, saving $180K annually.
Projects
Applied ML Research Projects
LLM Anomaly Detection
RAG pipeline using FAISS over 100K+ historical reports. Reduced hallucination rate by 40% through retrieval-based grounding. Findings documented in 25-page technical design adopted organization-wide.
THANOS Simulation Tool
Self-service platform for Middle East & North America supply chain optimization. Monte Carlo engine modeling 20+ variables projects $2M+ annual savings.
RAG Document QA System
FAISS + fine-tuned T5 pipeline for 10K+ technical documents. Achieved 87% accuracy through systematic experimentation with 5+ embedding models including Sentence-BERT and MPNet.
A/B Testing & Experimentation Excellence
Category Team Impact
Designed and executed A/B experimentation framework for retail promotions using causal inference methods. Analyzed 200K+ sellers and improved ad ROI by 12%, contributing to $2M annualized revenue.
  • Developed predictive impact estimation models
  • Reduced analysis turnaround time by 40%
  • Engineered automated ETL pipelines processing 5M+ daily events
  • Eliminated 70% of manual reporting workload
Education
Academic Background
Indian Institute of Technology (IIT) Madras
Post Graduate Program in Data Science & Artificial Intelligence
Mar 2022 – Mar 2023 | Chennai, India
Maulana Abul Kalam Azad University of Technology
Bachelor of Technology in Computer Science & Engineering
Aug 2012 – Jul 2016 | Kolkata, India
Professional Certifications
Machine Learning Specialization (Stanford/Coursera)
Deep Learning Specialization (DeepLearning.AI)
Google Data Analytics Professional Certificate
SQL Gold Badge (HackerRank)
Let's Connect
Applied Scientist passionate about building production ML systems that deliver measurable business impact. Open to opportunities in ML research, LLM applications, and scalable AI solutions.
Phone: +91 9433045127
Made with