Massachusetts Institute of Technology

Master's in Distributed ML @MIT '20  -  CS @MIT '19

Building Core ML infrastructure to empower Yelp-wide ML usage.

Yelp

Software Engineer, Core ML

Improved Yelp's photo classification service by 17% by re-writing and training the dish classification modelImproved coverage of classified photos by 33%. Improved model training pipeline with better frameworks + scripts.

Machine Learning Intern

Built a more robust and reliable mechanism for handling parameters in Yelp's internal experimentation platform. The project involved baking in type-support across the system. Stack - Python, JavaScript, React, and SQL Alchemy

Software Engineering Intern

Built linear models for confidential clause classificationImplemented the module for sub-clause parsing and analysis, improving the classification accuracy as much as 50% for each category compared to out of the box linear model.

Klarity

- Interned in backend software engineering team of a venture backed Pre-IPO healthcare-tech bay area         startup.•	Designed and developed Noho as company’s data platform- a petabyte scale data querying and         analytics platform built on top of AWS•	Wrote unit and functional test suites using mock and patch libraries for Noho•	Technologies involved – AWS services, S3, DynamoDB, Kinesis Streams, Lambda Functions           ElasticSearch, Kibana

HeartFlow, Inc

Working as an Undergraduate Researcher with Michael McDonald, Associate Professor of Physics at MIT, I worked to apply and optimize ASMOOTH, an adaptive smoothing algorithm by Ebeling, White and Rangarajan developed in 2006. I studied practical applications of ASMOOTH in visualizing X-Ray image data from distant massive galaxy clusters. The project involved a detailed study of the algorithm, its implementation for smoothing astronomical image data, optimizing free parameters for unambiguous visual structure and exploring mathematical methods to further improve algorithm’s output and performance.Technologies involved: Linux Environment, CIAO (X-Ray data analysis software), CSmooth, Python, Numpy and Scipy

Undergraduate Researcher

I initiated, developed and implemented a Science Summer Program for Indian high-school students by partnering with Avanti learning Centers. Our program focused on teaching the fundamentals of Computer Science, Robotics and Physics via a mix of hands-on activities and lectures. Students learnt about advanced Physics topics including Quantum Mechanics and Theory of Special and General Relativity, In the Robotics section, students built their own Arduino based Bluetooth controlled car.•Conceived the program as a means of fostering genuine interest in Science, inspiring students to recognize Science as an intellectual pursuit rather than merely a dull subject full of random facts and equations. •Worked directly with the CEO-- Akshay Saxena-- in order to shape the program and expand our reach to two different cities. •Created and structured the content of the program with significant input from my team and the CEO.•Organized and led a team of 5 enthusiastic undergraduates from top Indian colleges who helped with the content creation and facilitated sessions. •Delivered more than 20 lectures on topics including Algorithms, applications of Computer Science, theory of relativity, Quantum Mechanics, Computational Complexity theory, Cantor's theory of mathematical infinities and cardinality.

Summer Science Program Initiative

Founder

Spent two weeks in the summer of 2014 on Yale’s campus for an intense academic programcombining eclectic topics such as climate change, energy resources, and global health throughseminars, discussions and lectures by Yale’s faculty.Received interdisciplinary insights, global awareness and critical leadership and communication skills. Conducted team-research and proposed a feasible solution for Ocean Acidification--a dangerous phenomenon of continuous decrease of pH in Earth’s oceans, by the uptake of CO2 from the atmosphere.

Ayush Sharma