AMAR SRINIVAS
Fremont,CA
Summary
Seasoned Data Scientist and Gen AI Engineer with over 8 years of core programming, AI development, and Marketing Mix Modeling experience. Expert in Python, AI frameworks, Azure PaaS Services, and advanced statistical analysis for behavioral datasets. Proficient in developing AI-generated text prompts, fine-tuning Large Language Models (LLMs), and conducting comprehensive analytics. Driven by a passion for leveraging AI and statistical models to uncover deep insights into customer behavior and relevant KPIs.
Overview
4
4
years of professional experience
1
1
Certification
Work History
Senior Data Scientist
Certified Collectibles Group
2022.07 - 2023.02
- First Data Scientist hired at Certified Collectibles Group and spearheaded the automation of grading and validation of a variety of collectibles
- Created the grading model for coins and the data infrastructure pipeline to help process the imaging data automatically
- Technologies used: Python, AWS, PyTorch, perceptual hashing
- Created the model used for outlier detection using Local Outlier Factor methodology to help with new collectibles that our model had not dealt with before
- This allowed us to cast a wider net when dealing with different collectibles
- Technologies used: Python, Pandas, OpenCV, PyTorch, scikit-learn
- Developed the model for grading and evaluating currency and notes
- Structured the model application by breaking down the pipelines into DAGs (Directed Acyclic-Graphs)
- Implemented a Feature Bagging model for Outlier Detection which ran the Local Outlier Factor methods on multiple projections
- Used the combination of these results for improved detection qualities in higher dimensions
- Built a coins model to help deal with Bulk Grading, more specifically the American Silver Dollar Coin
- Had to create a model that accounts for various types of defects and paired with Elastic Search to use NLP to break down the verbiage available to us on the coin
- Developed a recommendation system to match collectible coins with potential buyers, leveraging machine learning techniques for personalization
- Technologies used: Python, AWS, Elasticsearch
- Work involved feature engineering, model optimization, and integrating the model into the existing pipeline using AWS and PyTorch
- Created models for outlier detection and grading currency, integrating these into Azure-based applications
- Developed AI-based recommendation systems, enhancing customer experience
- Contributed to the development of AI-generated text prompts for customer interactions.
Data Scientist/ML Engineer
Cedrus Digital
2019.04 - 2022.05
- Led the Cognitive Business Automation team at Cedrus in Data Science and Machine Learning development and delivered several successful customer projects
- Implemented Marketing Mix Models for retail clients to analyze and forecast sales, significantly boosting marketing effectiveness
- Conducted A/B testing and media mix analysis for a large-scale cosmetic brand, leading to a 30% increase in targeted campaign effectiveness
- Spearheaded the development of an advanced speech-to-text chatbot utilizing Python, NLP, NLG, and NLU technologies, revolutionizing user interactions within the application
- Orchestrated seamless data ingestion and real-time transcription using IBM Watson cloud services, enhancing the bot's ability to process spoken language, and converting it into accurate textual format
- Integrated cutting-edge Natural Language Understanding (NLU) techniques, supported by IBM Watson, to decode user intents and extract meaningful information from input messages, facilitating context-aware responses
- Implemented Azure PaaS Services for end-to-end Weekly Sales Retail Report generation
- Automated classification systems and developed a personalized chatbot experience
- Employed Natural Language Generation (NLG) methodologies, in conjunction with IBM Watson, to dynamically craft human-like responses tailored to individual user queries, elevating the overall user engagement and satisfaction
- Leveraged Python libraries including NLTK and spaCy, alongside IBM Watson NLU, for comprehensive keyword analysis, sentiment assessment, and language structure processing, contributing to an intelligent and intuitive chatbot interface
- Implemented a comprehensive and end-to-end Weekly Sales Retail Report using Timeseries forecasting for a large-scale cosmetics brand
- Technologies used: Python, AWS, SHAP (Shapley Additive exPlanations)
- Designed and implemented an Intelligent Character Recognition system with a major health insurance provider to accurately detect handwritten notes and checks
- Technologies used: Python, IBM Watson Discovery, Rasa NLU
- Automated a classification system for a medical private practice to properly classify in-vitro fertilizations using Microsoft Azure's computer vision platform
- Was able to improve classification metrics by 60%
- Technologies used: Python, Microsoft Azure, OpenCV
- Led the development of a personalized chatbot experience by integrating a recommendation system using NLP and machine learning algorithms
- Utilized IBM Watson and Python libraries like NLTK and spaCy for recommendation engine
- Managed end-to-end development, from data extraction and model training to feature development and deployment.
Data Science Fellow
Galvanize Inc
2018.09 - 2018.12
- Built a linear regression model to test and predict the sales value of tractors given various guidelines and parameters
- Ran cross-validation checks to pull the most pertinent models needed
- Created a Fantasy Football Player recommendation system as my Capstone Project
- Methodologies include Random Forest Regression and Gradient Boost.
Education
Extension - Visualization & Data Analysis in R -
UT
Austin, TX12.2023
Galvanize DS Immersive Program
San Francisco, CA12.2018
Bachelor of Science: Mathematics -
University of Texas At Austin
Austin, TX05.2017
Skills
- Python
- Webscraping/Bootstrapping
- Time Series/LSTM
- Dimensionality Reduction
- NLP/NLG/NLU
- Azure PaaS Services
- Machine Learning (Scikit-learn, Pandas, Numpy, PyTorch)
- SQL, PostgreSQL, MongoDB
- Scikit-learn, Pandas, Numpy
- AWS ML Services, IBM Watson, Microsoft Azure
- BluePrism Robotic Process Automation
- Deep Learning (TensorFlow, Keras)
- Marketing Mix Modeling, Econometric Modeling
- Statistics (Frequentist + Bayesian)
- Hypothesis testing, A/B Testing
- Visualization with matplotlib, seaborn
- Parallel Processing
- OpenCV
- Large Language Model Tuning, AI-Generated Text Prompt Development
- ROI Analysis and Optimization
- Media Spend Analysis
Certification
AWS Certified Machine Learning - Specialty, 2019
Timeline
Senior Data Scientist
Certified Collectibles Group
2022.07 - 2023.02
Data Scientist/ML Engineer
Cedrus Digital
2019.04 - 2022.05
Data Science Fellow
Galvanize Inc
2018.09 - 2018.12
Extension - Visualization & Data Analysis in R -
UT
Galvanize DS Immersive Program
Bachelor of Science: Mathematics -
University of Texas At Austin
AWS Certified Machine Learning - Specialty, 2019
Similar Profiles
ERIKA DILLONERIKA DILLON
Customer Service Representative at Certified Collectibles GroupCustomer Service Representative at Certified Collectibles Group
Norvik Morales CortezNorvik Morales Cortez
Pressing Comic Operator at Certified Collectibles GroupPressing Comic Operator at Certified Collectibles Group
Tebogo Amone RakgokongTebogo Amone Rakgokong
CEO at Certified GroupCEO at Certified Group