Saumaan Momin

Experience

• 

  Dell Technologies

  May 2019 - Nov 2019

  Data Engineer

  • Was part of CSAT team and aided in developing an application that automates ETL process.• Performed Data Migration from Amazon S3 to Amazon Redshift.• Worked on integrating the Hive data warehouse system with other data sourcessuch as Amazon Redshift.

• 

  EC Infosolutions Pvt Ltd

  Feb 2020 - May 2021

  Data Scientist

  • Assisted in scientific research on Animal Electrochemical Impedance Spectroscopy with smart barn and analyzed the results using advanced statical analysis and BI Tools. • With the help of object detection techniques (SSD and YOLO) utilize animal barn spacing. • Help in predicting and identifying the disease, pregnancy, ovulation, etc., in the animal at early stages with 82% accuracy. • Developed and implemented new models with statical reports which increased the research and productivity by 40%.• Conducted regression analysis(Time-series & Statistical modeling), achieving a 29% more accurate prediction of performance than previous models. • Utilized Web scraping techniques and API calls to get the data for forecasting. • Create a real-time pipeline on AWS-Sagemaker to automate the entire process, resulting in a 22% increase in usability for non-technical staff members. •Increased accessibility and usability of customer data by redesigning data visualization techniques to include statistical graphs and information graphics.• Develop custom shape predictor model from scratch to find and extract facial feature landmarks for dogs using Dlib, Opencv & Neural Networks (CNN) •Single Shot Detector(SSD), YOLO, and RCNN were used with the extracted facial feature to identify the dog’s emotions. •To support mobile devices in a real-time application, the entire prediction pipeline is converted into a TensorFlow Lite model.•Work on client's website & android application to deploy models with Rest API. Show less

• 

  Motilal Oswal Financial Services Ltd

  Jun 2021 - Sept 2021

  Senior Data Science Manager

  • Working with the Risk team to predict Cheque bounces and Credit behaviour over time using machine learning and Statical concepts• Created a rule engine using deep learning & graph analytics to identify when one will become non-performing assets & the reason behind them (causality)• Worked on the following services Azure Data Factory, Data Lake, Azure ML, MySQL Postgre, etc.

• 

  TomTom

  Jun 2021 - Jun 2023

  Data Scientist

  • Working with MCP-Engineering to manage the real-time MAPs data flow on Azure and uncover the insights.• Implemented the auto-serving of task model using Large Language Models (LLM) & Graph analytics for road edits. Which speeds up the process by more than 20x times while maintaining the same efficiency.• Created a 360 view of MAP editors to predict asynchronous wait time between the tasks. This alert increases productivity by 28%• Used Pysprak and Scala, to optimization of data pipelines on Azure• Worked on product and sales optimization algorithms using the NLP, location intelligence and statical analysis for o Sentiment in prospect communications o Red flags (like when a deal is in the final pipeline stage, but pricing hasn’t been discussed) o Sales emails, respond with a calendar invite to set up a sales demo. o Sales activity prioritization o Lead scoring o Customer lifetime value (LTV) modelling• Working on path routing operations in case of traffic, and road construction to provide an optimal path in real-time• Created a Customer Sentiment & Emotion Analysis model for product performance on various platforms & received the STAR AWARD (Best result)• Working on LLMs (BERT, BART, OpenAI's GPT-3/4, Facebook's OPT, and BigScience's Bloom) with the help of LangChain to automatically resolve customers', clients', and stakeholders' queries and issues.• With the help of Reporting team created Tableau dashboards for Business Stakeholders (Apple, Microsoft, Uber, etc.). Show less

• 

  Paytm

  Jun 2023 - now

  Senior Data Scientist

  • Led the Copilot project to automate the generation of monthly reports and ad-hoc analyses.• Developed an AI assistant utilizing multi-agent architecture and function calling to manage reporting and service desk tasks with AWS bedrock large language models (LLM) & advanced RAG pipeline(Hybrid search & Re-ranking).• Led the deployment, optimization, and monitoring of Large Language Models (LLMs) in production