Data Scientist Principal
Resume:
Ashburn, VA
:mehrdadshahabi mshahabi
:*******.*******@*****.***
Summary Principal Data Scientist / ML Engineer with 8 years of experience developing and de- ploying AI/ML solutions, specializing in Generative AI, NLP, fraud detection, and telemetrydataanalysis. Skilled in building and deploying GenAI models using LangChain within AWS SageMaker, and Databricks MLflow, leveraging CI/CD pipelines for seam- less production integration.
Industry Principal Data Scientist, Walmart Global Tech Jan -Present Experience - Building various Agentic AI Applications for detecting risk in Walmart stores
(deployed in LangGraph and Langchain)
Data Science Manager, GM Financial March -Dec
- Leading a team of four data scientists responsible for end-to-end delivery of various ML models using telemetry and customer data (deployed in Azure DataBricks through CI/CD pipeline).
- Analyzing various GM telematics data to build attributes for insurance pricing
- Designing a RAG based insurance agent chatbot that can interact with insurance claims data (using LangChain).
Applied Scientist: Buyer Risk Prevention (BRP), Amazon.com Jan -March
- Building various ML pipelines for fraud detection (using AWS SageMaker)
- Quantifying the Down Stream Impact (DSI) of incorrect decisions made by ML models with causal inference (using EconML Package)
- Fraud detection in customer contacts leveraging BERT (using HuggingFace)
- Real-time fraud detection using click-stream data (using AWS SageMaker) Data Scientist: AI/ML Team, Ford Motor Company Oct -Dec
- Using various supervised and unsupervised techniques for anomaly detection in time series (using PyTorch )
- Developing NLP based models for incorporating customers feedback into classification models (using PyTorch )
- Design and implementation of Multi-Arm Bandit methods for targeted advertising(using Python)
- Implementing various ML methods for predicting customers click through rate
(using Python, SparkML)
Skills Machine Learning Packages
- Python, Tensorflow, Keras, Pytorch, HuggingFace, scikit-learn, Catboost, XGboost
- Cloud: AWS, AWS SageMaker, AzureMl, Databricks, MLFlow Generative AI
- LangChain, LlamaIndex
Optimization
- CPLEX, GUROBI, IBM OPL, GAMS, Pyomo
mehrdad shahabi—r of
Academic Research Fellow: University of Michigan Experience Project: Design of Recommender Systems for Ride-Sharing Platforms: Ford Smart Mobility Analytics
- Design of an offline optimization based ride-sharing partner recommender systems(using Python)
- Ride-shared demand prediction using deep learning(using Python Tensorflow)
- Deep learning for trip energy estimation (Python TensorFlow)
- Developing distributed algorithms for robust data driven optimization problems Research Scientist: University of Florida Project: Design of back-end optimization algorithms for IncenTrip: US-DOE
- Design and implementation of a reinforcement learning approach for real-time incentive allocation problem (using Python)
- Developing innovative solution algorithms for solving real-time incentive allocation problems (deployed on AWS)
- Mining of 30+ Million trajectory data (Python)
Education West Virgina University, Morgantown, WV
Doctor of Philosophy, Transportation Engineering
- Thesis: Robust Shortest Path Problem; Formulation and Solution Algorithm Sharif University of Technology
Master of Science, Transportation Engineering
- Thesis: Day to Day Evolution of Traffic Flow Models Selected
Publications - An Integrated, Personalized, Real-Time Traveler Information and Incentive Technology for Energy Efficient Mobility Systems. Transportation Research Part C.
- An Outer Approximation Algorithm for Robust Shortest Path Problem, Transportation Research Part E: Logistics and Transportation Review Miscellaneous
- Coursera: Generative AI with DeepLearning.AI
- Kaggle Competition: IEEE-CIS Fraud Detection
- Devolopment and tuning of XGB methods for detecting fraudulent transactions
Contact this candidate