Darsh Shah

Darsh Shah

Software Engineer at Salesforce.com

Contact Me

About Me

I work as a machine learning engineer in the Search Relevancy team at Salesforce. I develop and test features and services that enable search relevance experimentations aimed at improving overall search ranking for Salesforce customers. Previously, I worked in search infrastructure team where I partly created and deployed a distributed system to manage thousands of search servers. I pursed my MS in Information Networking at Carnegie Mellon University. My focus of study was Distributed Systems. I am interested in understanding how current large-scale distributed systems are built and how to achieve scalability, caching and fault-tolerance in distributed systems. I have also have a certificate in Big Data technologies from University of Washington.

In my previous incarnation, I was a hardware tinkerer.

Previous Work Experience

Graduate Teaching Assistant - Carnegie Mellon University (2014 - 2015)

Teaching Assistant for ECE graduate level 18-641 Java for Smartphone Development course. Helping students in understanding the concepts of Java including design patterns and concurrency, Java Servlets/JSP and Android Programming.

System Software Intern - Apple Inc. (2014)

In the EFI(Firmware team), I was responsible for making infrastructure libraries to support diagnostics on production line. Implemented software solution which helped in increasing UPH (units/hr) during production

Associate Engineer - Qualcomm India (2012 - 2013)

Worked on FastCV, a computer vision library by Qualcomm. Wrote APIs in C and automation in Matlab and Perl to test Vuforia Augmented Reality App

More information on LinkedIn



Made a social network where people can add post, can add comments to posts and can follow/unfollow other people. Used HTML5/CSS3/JavaScript/jQuery for frontend. Django and Postgresql for backend. The application is hosted on Heroku.


Pittsburgh Transit Simulator

Made a Transit simulator application in Java which can simulate riders and buses in the City of Pittsburgh. The application can calculate delays by simulating rider density, weather conditions, different type of riders (wheelchair/luggage/group), payments methods (card/rfid/cash), type of buses and traffic for one whole day of operation.


Was a part of the team that made the website for Spacer, a startup idea in Hack-a-Startup competition organized by Tepper School of Business, CMU. I worked on the frontend and used HTML5/CSS3/JavaScript


Port Authority Realtime Application

Was a part of the team that created a web application which showed the realtime bus tracking information for Port Authority. It was a part of TartanHacks hackathon at CMU. I worked on the realtime API and backend part in Django.

Scrabble with stuff

Create a scrabble game in Java and implemented the GUI using Swing. The scrabble game has special invisible tiles that activate when a player plays a normal tile on it.

Search Engine Algorithms

Implemented a Boolean retrieval algorithm that evaluate queries and ranks documents based on Lucene. Implemented BM25 and Indri ranking along with query expansion to achieve better accuracy.

Input Text Predictor

Built an input text predictor, similar to Google instant, by generating a statistical language model on a list of n-grams generated from a text corpus using Hadoop Map Reduce in Java.

FitPlay: A Gamified Social Fitness App

Developed an android app which motivates people to do exercise using concepts of gamification. The app's backend service was hosted on AWS EC2.

More information on LinkedIn