Nov 2020 – Oct 2025 · 4 yrs 11 mos · Bengaluru, Karnataka, India

• Responsible for creating and managing the ML roadmap for flipkart wholesale.
• Worked on a variety of ML problems like cross cart recommendation, look alike modeling of customers, customer segmentation, search result ranking etc. for FKW.

Jul 2019 – Nov 2020 · 1 yr 4 mos · Delhi, India

• I am the Innovations Delivery Lead for Collab AI powered Interventions. In this role, I have worked on two projects
• Collab AI innovations for Lead to Cash (L2C). IBM is developing an Intelligent Workflows platform, and is onboarding L2C processes on it to start with. My team’s engagement is geared towards enabling Collections and Billing sub-process by deploying Social Machines on the platform. I contributed to designing Social Machines comprising of several agents for both sub-processes. I also designed some of the Agents (customer segmentation AI agent, AR Forecast AI agent) for Collections process. Presently, I am developing a Mechanism design AI Agent for the Collab AI platform.
• Contingent Labor Price Prediction (CLPP) for IBM procurement. The CLPP Social Machine consists of 4 agents - Naive, time series, Macroeconomic and Econometric. I contributed to designing the Social Machine and developed the Econometric Agent.

Jul 2014 – Jun 2019 · 4 yrs 11 mos · Delhi, India

• Jan 2017 - Oct 2019: I lead a research team to design and develop AI algorithms for IBM Watson Recruitment (IWR) job match functionality. The project has lead to conference papers, patents and an "A" level Research Accomplishment at IBM. The project involved several innovative and novel methods for feature engineering and data transformation to create AI models on past hiring data for deriving insights applicable to automated hiring analysis. The project involved:
• System and Methods to map any job to a standard job taxonomy (using NLP).
• End to end ML pipeline for training success models - Data cleaning, feature engineering
• and selection, model selection and hyperparameter optimization.
• Presenting outcome in an end user consumable manner.
• Boilerplate text identification in job descriptions.
• As part of my involvement with IBM Watson Talent, I also contributed to Watson Career Coach. I designed a TOPSIS and shortest path based methodology for recommending career progression plans to boost employee satisfaction and retention.
• Previously: Worked on an assortment of recruitment analytics problems creating demos, doing stand alone studies for various clients, and developing ML templates (using Watson Analytics) for a variety of scenarios. These contributions towards IBM’s global initiative of developing smarter workforce applications lead to an Outstanding Technical Achievement Award in 2016 at IBM. The experience gained in these projects culminated in conception of IWR.

Apr 2012 – Jun 2014 · 2 yrs 2 mos · Delhi, India

• Worked on Smarter Workforce Technologies for Contact Centers. This project enabled workforce management analysts to solution an account with
• Accurate workload forecasting
• Robust capacity planning for agents and seats
• Subsequent resilient scheduling of these resources to deliver services with optimum utilization and minimal delivery cost subject to other business constraints
• Simulation capability to predict SLAs for a given forecast and staffing plan.
• I joined the project in 2012, and lead it in 2013, resulting in $26M impact and an Outstanding Technical Achievement Award at IBM in Dec 2013. I developed the front and back office capacity planning engines (using IBM CPLEX), the simulation capability in Anylogic, and supervised onbaoarding of several client accounts.