Feb 2024 – Nov 2024 · 9 mos · San Francisco Bay Area · Hybrid

• Tech Lead and Architect for GenAI Workflows (LLM Agents and Tools) - specially MACAW (Multi Agent Conversational Agents Workflow).
• The GenAI Workflows platform enables Multi Agent experimentation, optimization, configuration and deployment ecosystem for complex workflows using planning, dynamic execution and in-context learning.

May 2023 – Feb 2024 · 9 mos · Menlo Park, California, United States · Hybrid

• Lead the fine tuning and development of several LLM based summarization, financial data extraction and relevance ranking pipelines.

Feb 2023 – May 2023 · 3 mos · San Francisco Bay Area · Hybrid

• LLM training and applications in Healthcare

Mar 2021 – May 2022 · 1 yr 2 mos · Mountain View, California, United States · Remote

• Developed adaptive learning and personalisation engines for one of the largest education technology companies in Indonesia. Was responsible for hiring a team of engineers and Data Scientists and develop strategies and vision to build products related to Personalized Learning.

Feb 2020 – Feb 2023 · 3 yrs · Palo Alto, California, United States · Hybrid

• Lead a team of ML engineers and Data Scientists at one of fastest growing conversational AI companies and developed products driven by Neural Search, Knowledge Discovery and Knowledge infused relevance ranking frameworks for automation products.
• Some of my achievements in this role include:
• 1. Develop a Knowledge Discovery module to reduce manual annotations needed to drive intent and entity extraction clients in diverse domains.
• 2. Developing a Domain Adaptation module for adapting language models for IT, HR and other domains.
• 3. Developed a Neural Search framework which powered search, QnA and recommendation systems
• across products in the company.
• 4. Develop, deploy and maintain the MLOps frameworks for a pipeline of NLP models.

Jun 2019 – Jun 2021 · 2 yrs · Mumbai, Maharashtra, India

• Lead a team of engineers and researchers to deliver several successful ML driven architectures for products in conversational AI, campaign optimisation, knowledge management and personalisation.

May 2016 – Jun 2019 · 3 yrs 1 mo · Mumbai, Maharashtra, India

• Lead the NLP and AI research and development for bot development, messaging app visualisation and recommendation systems for marketing campaigns. My work revolved around NLU, Semantic Search, Machine comprehension and Dialogue Modelling along with analytics and A/B testing.

Feb 2017 – Jan 2019 · 1 yr 11 mos · Delhi, India

• I acted as an advisor to the education and technology department at a grade A research institute of UNESCO to develop text, audio and video analytics software. Also developed an adaptive learning driven Learning Management System under the principles of "Collective Human Intelligence"

Jun 2015 – Jun 2016 · 1 yr · Mumbai Area, India

• I was responsible for building and training the machine learning team to execute products at various levels of the organisation's product stack. From analytics dashboards, to NLP and ML driven decision making, to AI systems augmenting the recommendations for the apps.
• As the product architect these were my responsibilities:
• 1. To build the content personalisation strategy.
• 2. To build prototype products which showcases how complex NLP (grammar induction) and machine learning algorithms (latent and graphical models) can be applied to the currently available data.
• 3. To build the data architecture and design the machine learning framework for execution of the recommendation algorithms in sync with the current Django based apps.
• 4. Design and develop the UI/UX experience on a front-end frame work like Angular/Ember.
• 5. Train and develop the content and curriculum experts to build content which can be fed into the adaptive system which requires minimum human intervention to work accurately.
• 6. Train and test predictive algorithms for intervention strategies.
• My skills lie in the following:
• Statistical and algorithmic machine learning - feature engineering/unsupervised learning, optimisation, neural networks (deep learning), hidden markov models.
• Natural language processing: Deep linguistic processing, natural language generation, semantic parsing and search, context inference, hybrid machine translation, information extraction (strings to relational databases, i.e. graphs) etc.
• Developmental cognitive neuroscience/psychology - functional and structural anatomy of the brain, numerical cognition, treatment of developmental disorders like dyscalculia etc.

Oct 2014 – Mar 2015 · 5 mos · Gurgaon, India

• As the product manager I developed digital and board based games for delivering a uniquely brilliant thinking skills curriculum to K-12 students. I developed a machine learning driven adaptive learning system which managed game progression for multiple games for a particular student and generated game levels according to a skill development based philosophy. My responsibilities included the following:
• 1. Develop algorithms for game generation and adaptivity according to the user level.
• 2. Map the dependencies between levels and learning paths for game players.
• 3. Develop cognitive testing systems using real time neuropsychological and cognitive tests.
• 4. Student profiling and reporting progress using data visualisation tools like d3js.
• 5. Managing the full stack engineers and integrate algorithms with the application stack to seamlessly build the adaptive platform.
• Technologies used:
• Front end - Angularjs.
• Visualization and game design - d3js, easeljs, pixijs.
• Backend - Nodejs
• Computational - Convnetjs and several graphical algorithms in nodejs and java.