Oct 2023 – Oct 2025 · 2 yrs · On-site

• Led engineering at ConverseNow.ai from pre-funding to Series A+, scaling conversational AI for 1,500+ restaurants and 1.5M monthly orders while managing 45 engineers across 3 continents.

May 2022 – Oct 2023 · 1 yr 5 mos · On-site

Oct 2018 – May 2022 · 3 yrs 7 mos · On-site

Aug 2017 – Sep 2018 · 1 yr 1 mo · Bengaluru, Karnataka, India

• Soroco is a process mining company that uses AI to help organizations understand and improve their business processes. I was hired as a report to the Head of Infrastructure and Security to build a team focused on improving the core platform microservices stack that serve as the backend of Soroco’s proprietary process mining software.
• I built and led a nine-member SRE team in Bangalore and my work focused on coordinating implementation and deployment of customized solutions for ML-based RPA.
• Project highlights included development of a patented algorithm to massively compress screenshots taken by a process mining application, leveraging image similarity models and video compression codecs (e.g. H265, AV1). The solution provides massive compression and storage savings with efficient single-read retrieval and high scalability, and it is used in production at several major clients.

Jul 2015 – Aug 2017 · 2 yrs 1 mo · Bengaluru, Karnataka, India

• I joined Samsung Research India Bangalore in a hands-on role focused on several deep learning solutions to optimize user interaction for a Conversational AI system. SVoice was an early Siri competitor developed for older Samsung models. Development included C++ algorithms for components related to multilingual named entity recognition (NER) and parsing.
• I was later assigned to Samsung’s Korea HQ for the initial design and development of the Domain Classification Model for Cross domain intent classification for Bixby 1.0, Samsung’s primary voice AI product used on all flagship smartphones and smart appliances. My contribution included development of the multi-level intent classification model using TensorFlow, model training, and data generation.