May 2024 – Aug 2024 · 3 mos · Boston, Massachusetts, United States · On-site

May 2023 – Aug 2023 · 3 mos · Cambridge, Massachusetts, United States · On-site

Aug 2022 – Present · 3 yrs 7 mos · Columbus, Ohio, United States

• Working under Prof. Huan Sun in creating a new benchmark for Information Seeking intents in Dialog Systems
• The proposed benchmark would test existing language models along multiple dimensions - open-domain QA, conversational QA, reasoning etc

Mar 2021 – Apr 2022 · 1 yr 1 mo · Bengaluru, Karnataka, India

• Worked on ranking & recommendation stack for real-estate search
• Implemented a low latency partial-query auto-suggest system incorporating textual and non-textual features
• Involved in Query Understanding & Categorization for free text queries
• Developed a similarity model for home listings for recommendation and comparative market analysis
• Worked on creating a unified representation using the multi-modal features on a real-estate listing page
• Developed a task-oriented chatbot to answer consumer queries in the absence of an agent - trained the dialogue state tracker to recognize the important entities and detect user intent as well as imbued the bot with a Reading Comprehension model to answer questions from text

Jan 2020 – Mar 2021 · 1 yr 2 mos · Bengaluru, Karnataka, India

• Leading a team working on applying and adapting novel dialogue state tracking systems and semantic matching techniques to improve sales conversations
• Scaling real-time deep learning services in production using multi-gpu inference and Kubernetes

Feb 2018 – Jan 2020 · 1 yr 11 mos · Hyderabad Area, India

• Mixed Objective, Block level optimization for web search recommendation in Bing using novel Pointing Decoder architectures - published a single-author paper on the same at SIGIR 2020
• Worked on Bing's domain-specific passage ranking and Machine Comprehension abilities - published the Attention-LSTM architecture through a paper at CIKM 2018
• Session-context aware and Multi-Task deep learning systems for follow up query suggestion for the Bing Search engine's Related Questions feature
• Novel contextual architectures for disambiguation tasks such as entity and word-sense disambiguation
• Collaborated closely with Microsoft Research, Redmond on various research initiatives in Dialogue Systems

Jun 2014 – Jan 2018 · 3 yrs 7 mos · Bengaluru Area, India

• Worked from ideation to implementation on several features towards the development of Amelia - a virtual assistant for BPO and help-desk automation
• Involved in the research and development of several modules including Dialogue Management, Machine Translation, Anaphora Resolution, Intent Detection and familiar with NLP tool-kits and frameworks including CoreNLP, Wordnet, Verbnet, Framenet, NLTK etc.
• Led a team working on improving state of the art approaches on tasks like Machine Comprehension, Natural Language Inference etc. using Attention-Matching Deep Neural Network architectures

Jul 2013 – Dec 2013 · 5 mos · Bengaluru Area, India

• Built an NLP module to check for any grammatical or spelling mistakes in text and returned a score based on the same. Integrated it with the companies existing servlet class for its simulations
• Implemented a Clustering algorithm to group individuals based on their scores when assessed across various skill-sets
• Developed an online assessment tool "Co{de}tective" for programmers and designed challenges to test the candidates in algorithms and data structures. Presented the same at the TechCrunch International Conference, Bangalore (2013)
• Created a 3D game in Unity3D as part of the company's internal hackathon using the Hofstede's cultural dimensions theory as the underlying theme. Experienced, first hand, the applications of game and decision theory involving complex inter-temporal choices

May 2012 – Jul 2012 · 2 mos · Goa, India

• Worked on digitizing the organization's inward-outward letter registers using ABAP programming on SAP ERP