Apr 2021 – Jan 2025 · 3 yrs 9 mos

• Building and managing a team of 30+ top quality Engineers and Data Scientists on a mission to build India's largest Hyperlocal marketplace.
• Overseeing the Product engineering process from Product Roadmap, Sprint planning, technical execution, delivery and Quality Assurance.
• Continuously measuring and improving the technical performance of the ecosystem, improving infrastructure stability and rationalising expenses.
• Hiring across roles in Product & Engineering. Checkout https://in.linkedin.com/company/lovelocalindia for open positions. DM dream@lovelocal.in for any enquiries.

Dec 2019 – Mar 2021 · 1 yr 3 mos

• Defined the initial set of core KPIs and OKRs across business and product to align with the business objectives and setting up monitoring systems to track them
• Built a team of high performing Data Analysts and Data Scientists to work on the crucial business problem statements and leverage data to come up with actionable insights
• Established rigorous sprint process in the Data Science vertical and breaking down of larger exploratory and experimental tasks into concrete deliverable chunks
• Built data products for retailers to make their digital store management smoother and derive business insights and recommendations from their LoveLocal listing
• Sett up data analytics platforms for business and product insights
• Developed Fraud Detection and management system to identify any behaviour of gaming the discounts and offers on the platforms by the network participant; reduced payment losses by 80%
• Developed Price, Product and Offer Recommendations for the retailers to keep their catalogue up to date as per latest consumer demand
• Developed a Personalisation module for the consumers to recommend products in their shopping journey, generate personalised nudges on repeating their purchases and curating a neighbourhood of shops
• Built Churn Prediction models to detect and flag to the account managers the stores that at a risk of churning in any given month
• Setup Geographical heat maps to understand and action upon the demand supply gaps for the Retailer and Consumer growth teams
• Sett up a Cataloguing Module to segregate the journeys for stores with a billing POS and stores without a billing POS, defining smart catalogues based on location, size and type of the stores.
• Developed an extraction, cleaning and processing layer to upload the data from
• Owned the entire retailer and consumer communications later right from defining engagement metrics, setting up communications using Push, SMS, WhatsApp and establishing the process to continuously optimise them

Sep 2018 – Dec 2019 · 1 yr 3 mos · Mumbai Metropolitan Region · On-site

• Setup the complete machine learning infrastructure building a separate micro-service to host ML models complete with a model management engine to tweak the decision boundaries of the model
• Built the data engineering pipeline to automate the periodic update of the models using freshly available data
• Integrated the risk models with the domain knowledge driven rule based decision system and introduced the process of probabilistic decision making in the organisation
• Built a graph database using connections extracted from contact list of users, developed algorithms to determine affinity between users and identify family members among connections
• Developed a one of its kind Graph based model to predict the risk associated with customers who have no history of credit, based on the embeddings derived from the credit graph. In the process, built one of the world's largest credit decisioning system that can predict credit risk on more than 500M users at that time.
• An application for patent for the above network based credit decisioning system is pending with the US Patent and Trademark Office
• Developed the decisioning system for Top Up loans and credit limit upgrades for users who have taken their first loan; collaborating with the Engineering and finance teams to launch the products
• Developed several prototype Machine Learning and Deep Learning models to consistently optimise the process of reviewing the loan applications from Image processing and validity check, address decoding, PDF parsing etc

Nov 2016 – Sep 2018 · 1 yr 10 mos · Mumbai Metropolitan Region · On-site

• Built ML and Heuristic Models to assess the credit risk of unbanked users
• Developed parsers on user SMS, Bank account statements, location and device usage patterns to profile users which would be fed to the Credit Decision engine. This is used to estimate the character, capacity and collectibility of the users when making the credit decision
• Built a lead prioritisation system for the sales teams to focus on high quality leads increasing funnel conversion
• Setup the entire credit and Business analytics dashboards and automated reports to monitor all key metrics and OKRs, portfolio health and vintages, delinquency levels and revenue performance.
• Developed an ML model to predict the collectibility of the users every month by reading signals from their SMS, credit reports, location and app usage patterns.

Jan 2016 – Nov 2016 · 10 mos · Greater Delhi Area · On-site

• An Online doctor consultation platform where patients post medical queries to a forum or a private conversation
• Developed Machine Learning models to classify the incoming queries from users into medical topics and then match the queries to the relevant doctors based on their affinity scores computed from the previous responses and preferences
• Semantic understanding on the queries to generate automated responses
• Built a model to monitor anonymised private conversations for any medical prescriptions that deviate from the patterns
• Developed a Spam filtering model to improve the quality of the queries doctors and users see on the platform

Mar 2015 – Nov 2015 · 8 mos · Powai, Mumbai · On-site

• Building In-house data products to assist the Business teams in smoother operations
• Developed an audit system to identify fraud in the collection of housing inventory data. The system identifies duplicate flats (detected by image matching), erroneous data, non-existent inventory listed
• Developed framework to track performance of the online Real-Estate market by scraping inventory across competitor platforms. Setup analytics to track the firm’s coverage and alerts to data collectors on availability of new inventory
• Worked on several exploratory analysis to quantify the demand-supply gap on the platform and setup processes with the acquisition teams to cover for the missing supply

Nov 2013 – Mar 2015 · 1 yr 4 mos · Mumbai Metropolitan Region

• Currently acquired by Fractal Analytics
• Independently executed several projects with multinational firms from diverse fields, like Ecommerce, Telecommunications, Agriculture, Food Technology etc
• Developed a global food database for FoodPanda by setting up an ETL process for the restaurant menus in 40+ countries. The pipeline includes extraction of data from from files, translation from the native language, spellcheck, identifying cuisine, category and ingredients and creating a tag cloud for easier Search and discovery.
• Created detailed profiles using geolocation information of users, determining the home and work location, travel and shopping patterns of a user etc.
• Scraped inventory data of leading E-Commerce sites; identified similar products, comparing popularity, pricing and available offers across the portals to setup competition benchmarking for Myntra and Jabong
• Worked with a P2P platform (Prosper.com) in their early years to setup their analytics and reporting systems. Developed the prosper score, a metric to measure the loan listing quality and built tracking systems to measure the performance of the score. Setup the ETL pipelines to automate the reports and refresh the score on schedule.
• Worked On-site as Data Science consultant for Monsanto to manage and improve their ML models for determining the crop life expectancy.
• Also worked as on-site technical consultant for Payback for the SAS based Marketing Automation platform