A. Introduction to Statistics

B. Mathematics for Machine Learning

Bachelor of Engineering - BE

12th

10th CBSE

Object detection in satellite imagery

•  Implemented a complete workflow for pole detection using satellite data and shapefiles.
•  Developed a POC on tree count & tree canopy detection using computer vision techniques.
•  Built an end-to-end training workflow for model onboarding on AWS. Executed data wrangling, pre-processing, model building and validation.
•  Handled raster data, shapefiles and street view images.
•  Worked on deployment practices like containerisation, logging & tracking.

Data Science Intern

Central Management System for Electric Vehicle Chargers with Load Forecasting

•  Designed a system for station and user management where user can connect to any charging station by locating it on maps, connect to charger to charge vehicle and pay online.
• Worked on developing charging & transaction modules.
•  Also implemented SOTA time-series forecasting machine learning models like XGBoost, SARIMAX and Facebook Prophet to predict load on any specific charging station.
•  Carried out research work to gauge potential in load forecasting using deep neural networks, analyse demand to manage supply-side and importance of smart-grids.

Project Intern

Coordinator

•  Engaged in the development and architectural design of the AIMS (Asset Inspection & Monitoring System) product, where I deployed an end-to-end automated & scaled workflow featuring CI/CD, statistical & machine learning solutions. Volume processed is ~130TB comprising of 80million images with an accuracy of 93% in capturing assets/defects.
•  Worked on multiple products for developing POCs on image-segmentation, object-detection, time-series, automation with SDKs & A/B testing of tools.
•  Led an effort to boost operational efficiency through optimization of memory usage via multiprocessing, refinement of Docker images, and implementation of data compression. This resulted in realized savings of approximately $XX,000 per month.
•  Constructing an automated asset-image mapping, distance estimation and geo-tagging algorithm using vision transformers & foundational models.
•  Took charge of defining project requirements and leading the in-house development of a labelling tool with AI-ca

Data Scientist

Data Scientist @AiDash | PICT '22

Know about Mihir Kulkarni's work experience, education, bootcamps, certifications, and much more. Mihir Kulkarni