Mar 2021 – Nov 2022 · 1 yr 8 mos

• # Leading research and development for Haptik's Conversational AI platform which powers 400+ Virtual Assistants, has reached 500 million+ consumers, supported 4 billion+ conversations and is growing exponentially.
• # Building the most sustainable ecosystem for virtual assistants which enables them become smarter, faster and cheaper with time
• # Built an advanced NLU that leverages LSTMs and Transformers (BERT, T5, GPT) to apply key concepts such as transfer learning, data augmentation, protection against adversarial queries. It works like Auto-ML service for building virtual assistants.
• # Enabled scaling to support 500+ training and 2 million+ inference calls daily with dynamic deployments and routing. We use Istio and Kubernetes to achieve isolation for multi-tenant machine learning model deployments.
• # Build a culture that encourages innovation and creativity and also aligns with Vision and Mission of the organization.

Apr 2019 – Mar 2021 · 1 yr 11 mos

• # Having joined back in 2014 as first Machine Learning Scientist on technology team, I set up Natural Language Processing stack of Haptik's conversational AI platform ground up and helped grow engineering team from 5 members to 75+ members.
• # Contribute to setting vision and strategy for machine learning models which supports Haptik's conversational AI platform.
• # Responsible for OKR planning and ownership deliverables of Machine Learning Team.
• # Mentor researchers and engineers on Haptik's ML team
• # Continuous experimentation to push accuracy and performance of short text similarity, text classification, answerability, named entity recognition and dialogue manager components
• # Created a reputed engineering brand by publishing blogs, papers, code samples and successfully attracted talent from Fortune 500 and Big tech companies
• Haptik is a part of the Reliance group of companies, who acquired a majority stake in the company in a $100 million deal in April 2019.

Mar 2016 – Mar 2019 · 3 yrs

• Having started ground up in 2014 as first Machine Learning Scientist at Haptik, I was incharge of building core machine intelligence of Haptik's Chatbot Platform. Following are few highlights of my role.
• # Lead research and development function at Haptik and own underlying NLP modules and Machine Learning pipeline backing Haptik's Chatbot Platform.
• # Proficiently implemented several rule based, conventional machine learning and deep learning algorithms and scaled them across multiple domains.
• # Coordinated with front-end, back-end and dev-ops teams to set up right infrastructure to process more than 5 millions messages everyday with responses time lesser than 1s.
• # Developed Recurrent Neural Network based question answering bot and put it in production for limited domains. You can find more details on my blog at https://haptik.ai/tech/neural-conversation-model-production/
• # Understand Product and Business requirements to drive Machine Learning effort of the company in result oriented direction.
• # Cut down operational cost of company by more than 65 percent.

Dec 2014 – Mar 2016 · 1 yr 3 mos

• # Joined as 3rd member on Haptik's engineering team and first one incharge to build Machine Learning pipeline from scratch
• # Designed various Natural Language Processing components such as Intent Detection, Entity Recognition and Information Retrieval to respond to text messages received on Haptik Platform

Jun 2018 – Dec 2022 · 4 yrs 6 mos · Mumbai Metropolitan Region

• Google Developers Experts (GDEs) are experienced, recognized developers of Google technologies as well as outstanding professionals in product strategy, UX/UI, marketing, growth hacking and monetization. They distinguish themselves through frequently speaking at conferences, share their passion and experience by publishing videos and tutorials, writing code samples, mentoring developers and startups and much more. Thanks to their support, developers, high-potential startups and technical communities around the world build and launch highly innovative apps.
• https://developers.google.com/community/experts/directory/profile/profile-krupal_modi

Jul 2013 – Dec 2014 · 1 yr 5 mos · Bengaluru, Karnataka, India

• # Designed and implemented customised algorithms for sentiment analysis, clustering and classification for social media analysis of companies across multiple domains.
• # Designed and implemented pattern recognition and image processing algorithmic flow for In-store Video Analytics of one of the leading retailers. Customer and staff behavior was analysed across the store using existing cctv infrastructure
• # Hands on experience at al stages of the project including data ingestion, processing and visualisation

Jan 2013 – Jun 2013 · 5 mos · Bengaluru, Karnataka, India

• # Implemented Pattern Recognition based Human Augmented Machine Learning algorithm for efficient Optical Character Recognition
• # Designed Crowd-sourcing and Evaluation workflow for augmenting human effort with machine learning algorithm for optical character recognition.