Experience

MOEV Inc.

• United States
• Software Development
• 1 - 100 Employee

• Senior Software Engineer, Machine Learning

  • Aug 2022 - Present

  Los Angeles, California, United States • Led the machine learning team in predicting the estimated time of arrival (ETA) in lastmile delivery, driver behavior and extracting information from GPS noise . This achieved less than a 5% route planning error.” • Implemented CI/CD pipelines for deep learning models, reducing route planning error by 40% in 6 months, enabling automatic model drift detection, and reducing model update time from 4 weeks to 2 days. • Developed low latency REST API’s and high-scalability ML production… Show more • Led the machine learning team in predicting the estimated time of arrival (ETA) in lastmile delivery, driver behavior and extracting information from GPS noise . This achieved less than a 5% route planning error.” • Implemented CI/CD pipelines for deep learning models, reducing route planning error by 40% in 6 months, enabling automatic model drift detection, and reducing model update time from 4 weeks to 2 days. • Developed low latency REST API’s and high-scalability ML production systems on AWS Serverless backend on our flagship ML models to promote secure data exchange with frontend.This led to a 60% reduction in API response time and we were able to scale up to Million calls a day. Show less



Iambic Therapeutics

• United States
• Biotechnology Research
• 1 - 100 Employee

• Machine Learning Scientist

  • Mar 2021 - Jun 2022

  California, United States



Varnerlab

• United States
• Research

• Graduate Deep Learning Researcher

  • Aug 2019 - Feb 2021

  Ithaca, New York, United States - With a 73% accuracy, assisted chemists to identify experimentally viable valid pathways using Monte Carlo Tree search for searching in 2048 branching factor state to make drug molecules. - Wrangled 1TB of doctors text notes from the MIMIC data using PySpark-SQL. Removed noise and mislabelled data using Autoencoders; GAN’s to fill the missing values; prediction was done using SVM.



Process-Energy-Environmental Systems Engineering (PEESE) Lab

• United States
• Research Services
• 1 - 100 Employee

• Graduate Researcher

  • Jan 2019 - Dec 2019

  Cornell University - Scraped real time weather data, solar and wind energy generation using Beautiful Soup. Worked on a MINLP workflow to track the optimal times to shift energy usage away from fossil fuels to renewable energy sources. - Worked with Autoencoders to track drops in data quality i.e. anomalies/noise in time series data and use missing value imputation to supplant them.



One Concern

• United States
• Software Development
• 1 - 100 Employee

• Data Scientist

  • May 2019 - Aug 2019

  Menlo Park, California Reduced inconsistency between Microsoft Building Footprints and Corelogic tax collection data by setting test cases to identify and remove erroneous data entries to improve existing ML models.



Govt Medical College

• Data Science Researcher

  • Jul 2017 - Jun 2018

  Srinagar, India - Early detection of patient developing High/Low risk of heart blockage (Atherosclerosis) by using SVM on sparse patient data; self-collected over 6 months study on 83 patients with an accuracy of 78%. – Extracted data from doctors notes, videos to quantify the percentage/presence of heart blockage. Correlated impact of long-term exposure to PM 2.5 air pollutants and heart health of population living near this industrial center.



One Concern

• United States
• Software Development
• 1 - 100 Employee

• Data Scientist

  • Jun 2016 - Sep 2016

  Palo Alto, California Estimated post-earthquake impact on different utilities of a city by implementing a web scraping tool from Twitter, Instagram with Beautiful Soup, Selenium and SQL.