Headline

Data Science at Nykaa | Ex-Swiggy | M.Tech-AI IISc Bangalore

About

Data Scientist at Swiggy. Trying to solve problems across domains like recommendations, customer profiling and discounting in the e-commerce industry.

Location

Bengaluru

AI stack

claude code

codex

gemini

kiro

langchain

huggingface

Work experience

Staff Data Scientist at Nykaa, Apr 2026 - Present · 6 mos, Bengaluru

Data Scientist III at Nykaa, May 2024 - Mar 2026 · 1 yr 11 mos, Bengaluru, - Churn Prediction Framework to identify the right time to act on users at risk of churn. - Customer State Transition Models to predict the movement of users from one affluence state to another. - Search - Working on revamping the Query Understanding pipeline to improve candidate generation - Generated product attributes for improving search using Gemini and GPT models.

Data Scientist II at Swiggy, Nov 2020 - May 2024 · 3 yrs 7 mos, India, - Formulated POCs for identifying and ranking for explore and exploit intents in listing. - Formulated a novel MAB based solution for ranking of widgets in Swiggy's new homepage * Objectives: addressing relevance, combinations and diversity of the page * Worked closely with the engineering team while leading the implementation efforts. - Trust and Safety - Currently own the Cash on Delivery abuse reduction efforts from DS end. * Working closely with business, product, engineering, finance and analytics stakeholders to drive multiple projects end-to-end * Working with highly unbalanced classes and low volumes of reliable data and rapidly evolving customer behavior * Built solutions for dynamic cart value limits for offering COD option to maximise conversion while keeping cancellation risk in check using uplift modelling. Published a paper on this subject in ECML PKDD 2023. * Coordinated session attribute generation across 6 engineering teams. These session attributes are used to detect anomalous sessions especially among new users. * Built a framework to optimise the cash carrying capacity of delivery executives - Helped optimise the Instamart hub locations using facility location formulation. - Hackathon - First prize winner in an internal hackathon among Swiggy Data Science and Applied Research team * Built an end-to-end flow for a cart-building voice assistant in Swiggy Instamart - objective was to deliver the kirana store experience for users * Built or used open source components wherever possible for modules like ASR, NER, TTS, Search and Ranking and tied them all together to give a hassle-free flow over a span of 3 days with 4 team members.

Data Scientist at Swiggy, Nov 2018 - Oct 2020 · 2 yrs, Bengaluru, Karnataka, India, - Built domain knowledge inspired solutions for Customer Churn prediction - Worked on the early stages of building a robust experimentation platform for Swiggy - Revamped the listing recommendations charter * Formulation and modeling of listing as point-wise classification, and later other LTR methods * Simple to complex solutions for multiple business objectives in listing - conversion vs delivery cost, different levels of funnel conversion (menu conversions vs cart conversions - MTL), cold start, etc. * Working on complex engineering challenges like huge data volumes, cost-effective inference strategies, monitoring and alerts on business and model metrics, and RCAs on escalations. * Delivered significant business impact and DS improvement over this period.

Business Analyst at Swiggy, Jun 2018 - Oct 2018 · 5 mos, Bengaluru, Karnataka, India, - Built different customer segments to help the growth team run targeted ads on Facebook - Studied discounting with a PhD Scholar and worked on Uplift Modelling to identify the right users to offer discounts and cut down discounting costs.

Data Science Intern at Swiggy, Jul 2017 - Dec 2017 · 6 mos, Bengaluru Area, India, - Deliver useful business insights around customer ordering patterns - Built ETA prediction models using different boosting and bagging methods - Built a simulation tool using Anylogic to visualise the order assignment logic in action - Built a simulation tool in Python to visualise traffic at different stages in users' app journey modelled using State Transition Matrix

Advisor at Enactus, May 2016 - Mar 2017 · 11 mos, BITS Pilani Goa Campus

Head, Workshops and CSR at Quark 2017, Apr 2016 - Mar 2017 · 1 yr, Goa, India., Quark has been a wonderful experience in the past two years. Teamwork and leadership are the core values that I learnt here. In my role as Head, Workshops and CSR I had the following accomplishments: • Managed a team of over 60 students. • Increased revenue while decreasing expenditure substantially compared to previous editions • Improved workshop quality by bringing companies in like Mathworks, TI, NVIDIA, Zerodha etc. to deliver workshops. • Organised multiple nation-wide Blood Donation-cum-Organ Donation - RED campaigns across India. • Secured UNESCO Patronage for the festival 'whY Chromosome' - a Women Empowerment Campaign.

Education

Master's degree, Artificial Intelligence, Indian Institute of Science (IISc)

Bachelor's Degree, Electrical and Electronics Engineering, Birla Institute of Technology and Science, Pilani - Goa Campus