Feb 2024 – Present · 2 yrs 3 mos

• I typically invest in YC companies

Jan 2023 – Mar 2024 · 1 yr 2 mos · Atlanta, Georgia, United States · Remote

Aug 2021 – Jan 2023 · 1 yr 5 mos · Atlanta, Georgia, United States

• (Homesite acquired by American Family Insurance)
• Intrapreneur award winner 🏆 for leading several key ML projects
• 1. Spearheaded the development of end-to-end ML pipeline to predict retention elasticity among customers. Leveraged Cloud Run and Kubernetes in GCP to build and enhance AI-driven tools to enable better decision making, achieve KPIs by delivering better data products.
• 2. Led data scientists and data engineers to design a ML pipeline to monitor the outliers in several key business metrics by leveraging Docker and Apache Airflow.
• 3. Development of AI based app using Streamlit to visualize data and steer effective communication with the stakeholders. This has been a key intrapreneural effort that currently supports many end users including product managers.
• 4. Currently involved in identifying key metrics to monitor for budget and planning data for all operating companies, partners and channels.

Oct 2018 – Aug 2021 · 2 yrs 10 mos · Houston, Texas, United States

• 1. Research and development of unsupervised, deep-learning based, image segmentation and analysis models. Build pipelines for fast and accurate interpretation of real world complex data. Patent filed.
• 2. Developed deep learning models for detecting and selective filtering of coherent noise from image and time-series data. Achieved 98% performance accuracy and deployed containerized models into the production environment.
• 3. Built improved predictive models using Python and R for several domain-specific big data problems.
• 4. Created powerful visualization using Power BI and Tableau to communicate data insights to business clients.
• 5. Frequently collaborated with product managers, software teams, and other scientists to ensure product success.
• 6. Actively provided practical ML solutions to customers and internal teams. Help customers deploy their models in production environment. Research and evaluate state-of-the-art ML methods.

Jun 2018 – Jul 2018 · 1 mo · Silicon Valley

• (1) Built a Natural Language Processing pipeline for document classification, for the task of fake news detection, by analyzing the incongruity between the headline-body pairs of news articles.
• (2) Developed and trained GRU and LSTM models that achieved 97% accuracy.
• (3) Served the model through an API deployed using Flask.

Oct 2017 – Jun 2018 · 8 mos · Greater Seattle Area

• Numerical simulation and optimization of space probe landing on Europa’s rocky surface. The modeling efforts are centered about two modeling formulations: smoothed particle hydrodynamics (SPH) and the arbitrary Lagrangian-Eulerian (ALE) set of techniques.

Sep 2017 – Aug 2018 · 11 mos · Greater Seattle Area

• Projects in Machine Learning
• 1. Object detection, localization, and retrieval of nearest neighbor with LSH - Used hard negative mining to handle class imbalance on MS-COCO dataset during object localization and also implemented Stochastic Dual Coordinate Ascent method for attaining faster optimization by 10%.
• 2. Sentiment analysis on amazon food reviews- End to end implementation and deployment of a web application that classifies sentiments based on food reviews. The application was created with flask and deployed on the Heroku platform. Used pretrained word embedding and NLTK package to build the LSTM based pipeline.
• 3. Exploring the night sky - Built an automated multi-class classification pipeline of detecting astronomical objects by using photometric redshift data from Sloan Digital Sky Survey (SDSS) .

Jun 2017 – Sep 2017 · 3 mos · Greater Houston

• Developed supervised machine learning models for multidimensional data analysis framework, with application to predicting subsurface parameters from seismic datasets. This is a
• patentable productivity enhancement technology

May 2016 – Aug 2016 · 3 mos · Greater Houston

• Software development for building a highly efficient parallel framework for cluster optimization
• and handling big datasets, for advanced data processing of oil exploration datasets.

Aug 2014 – Apr 2017 · 2 yrs 8 mos

• (1) Developed dimensionality reduction and optimization algorithms.
• (2) Research paper presented at SEG conference.
• (3) Designed and built high performance software modules for wave propagation in complex media.

Jun 2013 – Jul 2013 · 1 mo · Mumbai Metropolitan Region

• Built neural network models for pattern recognition and feature characterization in high
• dimensional seismic images, outperforming existing algorithms.