I work as a Principal AI/ML Engineer at Twilio in the San Francisco Bay Area. In my current role, I'm primarily responsible for developing Twilio's GenAI capabilities. Previously, I've worked at SAP Labs in Palo Alto and later at Newport Beach, CA. Most recently, I've worked at a startup called Zaplabs (acquired by Anywhere RE). I graduated with my master’s degree in Computer Science from Indiana University in Bloomington with a focus in the areas of Machine Learning and Data Science. Prior to my graduate studies, I was as a Research Assistant in the Dept. of Computation & Data Sciences at the Indian Institute of Science in Bengaluru, India.
San Francisco, CA | Feb ‘23 - Present
I'm building GenAI powered applications and tools for customers of Twilio & internal stakeholders (in addition to some classical ML propensity models for our GTM teams). Working alongside Zach Merritt has been great and we've been celebrating each little victory with a well-timed virtual high five!
Emeryville, CA | Oct ‘19 - Feb ‘23
I was involved in end-to-end development & deployment of Machine Learning models for a line of products offered by Zaplabs (acquired by Anywhere RE) in the Real Estate space. I was lucky to find an awesome mentor in Tony Kueh here.
Palo Alto & Newport Beach, CA| Sep‘17 - Oct‘19
I worked as a Data Scientist in the SAP Cloud Machine Learning team and was responsible for building ML use cases and deploying them into production environments. I had the opportunity to learn from and work with brilliant engineers like Peter Colligan.
Indianapolis, IN | Jan ‘17 - May ‘17
I interned with the Obamacare Infrastructure & Analytics team. I was taught a great deal about Data Infrastructure, U.S. Polity, and cars by Jon Walthour.
Bloomongton, IN | 2015-2017
I went to school here. I studied, researched and helped teach. IU is ranked among the most beautiful campuses in the country. You would’ve almost always found me in one of these places on campus. I was mentored by the amazing Dr. Travis Brown.
Bloomington, IN | 2016
I interned here for the summer, helping build .NET web and cross platform mobile applications for an enterprise facing line of technology products by Bloom. I had a great summer working with Derek Foreman.
Bengaluru, India | 2014-2015
I did research in parallel computing for Intel. Being India’s #1 University, IISc is known for it’s amazing campus. I learnt a whole lot sharing the lab with smarts like Prakash Murali & Nitin Bhat.
Bengaluru, India | 2014
I interned here during the summer with the R&D team, working with Saurabh Sood to develop hybrid mobile applications for Unisys.
Bengaluru, India | 2013
I interned with General Electric's software org during my senior year and helped streamline multiple modules of the CRM, optimizing queries and standardizing logging practices.
Twilio
Developed an LLM powered internal Q&A app with up-to-date product information, technical specs, customer stories, CRM data, sales playbooks, prescribed strategies and transcribed product demos. This app serves as an AI assistant for the GTM team, aiding in personalized outreach, looking up product info, code generation, competitive intelligence, and market research.
Twilio
Created an automated RFP (Request For Proposal) response tool that utilizes LLMs with first party product & technical information, streamlining the response process through parallelization of API calls and batching. This helped cut our response times from about 3-5 weeks (with multiple sales personnel involved) to < 15 minutes. The quality of the responses was manually analyzed, with > 90% acceptance rate.
Zaplabs
Developed and deployed models to identify top buy & sell side leads for real estate agents and brokerages. Used across multiple Anywhere RE brands (Coldwell Banker, Century 21, Sotheby's etc.) to generate monthly leads for 4000+ brokerages.
Zaplabs
Recommendation model using Network Graphs to identify key real estate agents and brokerage offices for referrals. Actively used in all Anywhere Real Estate Brands for agent referrals. Saw > 30% increase in success rate of referred closings.
Zaplabs
Time Series Forecast model to predict future market share company owned real estate brands, franchisees, and competitors across varying perdiods.
SAP
Predictive model built into SAP Service Cloud to estimate the time it would take to address customer service tickets.
SAP
A constraint optimization use case developed for SAP Service Cloud to auto assign leads to the most relevant sales representatives while optimizing on a number of key metrics.
SAP
Developed an automated model deployment pipeline for SAP Cloud Platform which was responsible for pushing models to production. Implemented TensorFlow Serving based deployment and developed GDPR compliant data ingestion pipelines. Ported ‘Boom‘, a Go based based load testing application to Python to integrate with deployment pipeline.
Indiana University
I was a part of Prof. Mehmet Dalkilic‘s research group and was involved in building custom tree based ensemble models in an attempt to collate known gene lesions responsible for human inherited disease.
Indiana University
I had the opportunity to assist Prof. Chanh Kieu in his research and was responsible for building and parallelizing atmospheric modeling systems (CAM 5.3) on the IU’s supercomputers – Big Red II and Karst.
Intel | Indian Institute of Science
I worked at Prof. Sathish Vadhiyar‘s lab at SERC (Now, CDS), IISc – building, optimizing, and executing parallel climate models of CESM on Intel’s MIC Architecture as a part of the Intel Parallel Computing Center (IPCC).
Twilio Comms Hackathon, 2023
Developed an LLM powered chat application with embeddings of all of Twilio's developer documetation to serve as a coding assistant for developers.
Anywhere RE Hackathon Winner, 2022
Developed a GPT powered self-serve application for agents to generate professional listings descriptions, SEOptimized ads, professionalized customer interactions and ability to query internal documents in natural language to find answers.
SAP Accessibility Hackathon, 2019
A Gesture to Speech module built using Alexa skills and Tensorflow to enable people who cannot speak to interact seamlessly with digital assistants in an enterprise setting.
SAP Hackathon Winner, 2018
Developed Tensorflow powered Text Summarization solution to condense large bodies of text. This was based on Google’s TextSum.
Indiana University Hackathon, 2017
Built a Jarvis inspired Twitter bot to generate random tweets based on popular content on Twitter using Markov Chains. It may sometimes seem to be surprisingly philosophical.
Unisys Cloud Hackathon Winner, 2014
Developed bio-metric authenticated payment solution. Won the all India 2nd place at Unisys Cloud 20/20 Hackathon. This project was further funded by the Dept. of Science & Technology, Govt. of India.
Consutant
Supporting Govt. of Karnataka (Office of the Addn. Chief Secretary) achieve the envisioned Sustainability Development Goals as a pro bono consultant for Data Science, Machine Learning & Analytics focused initiatives.
Volunteer
Volunteering to help the tech team at Isha Foundation with GenAI efforts.
Mentor
I've been invited to speak to the graduating batches of the Data Science program at Notre Dame, and have been involved in mentoring folks from the cohorts.
I appreciate logic, wit and humor. Topics which I closely follow are India, Technology, Education, Yoga & Clean Energy. I blog occasionally and take to Twitter to share things I find interesting. You can chat with my digital persona below.
My Myers-Briggs Type Indicator is ENFJ-A. Apparently, I share my personality type with the likes of Nelson Mandela, Barack Obama & Sheryl Sandberg. You too can feel good about your personality type here.
I like spending my time tinkering with hobby projects, taking walks with my wife, catching up with friends, playing racquetball, swimming, and making chai. I am on a lifelong quest to learn new languages. Lately, I’ve become interested in learning more about finance & public policy. Here is a collection of my favorite stuff that I enjoy watching, listening to or reading.