Jun 2020 – Aug 2023 · 3 yrs 2 mos · Menlo Park, California, United States

Apr 2018 – Jun 2019 · 1 yr 2 mos · Bengaluru Area, India

• To ensure customer safety at Ola, developed an algorithm to generate ride safety scores in real-time(​in production​). Generated booking hotspots and spatial clusters using agglomerative clustering. Utilized linear models & anomaly detection techniques to generate safety index for all possible routes between clusters using gps signals, historical rides, traffic density, complaints and reported crime data.
• Detected potholes with 86% accuracy using signal processing and clustering techniques on the data from accelerometer sensors. Generated monthly HTML report with pothole location markers on google map and shared with the state ​government​ for action.
• Developed a logistic regression model to generate driver scores based on their driving behavior(harsh acceleration, brakes, speeding etc.), customer complaints, past accidents and vehicle maintenance records - convinced operations team to use it for driver training.
• Technical skills: ​PySpark, Python, Hive, Qubole, Kibana, Microstrategy, Git, AWS, Scikit-learn, Matplotlib, Dlib, OpenCV, Leaflet

Apr 2017 – Mar 2018 · 11 mos · Bengaluru Area, India

• Analyzed commute patterns of users, predicted their home and work locations, identified routes with highest demand density and used this information in supply shaping which resulted in a reduction of demand-supply gap by 1.5%.
• Designed experiments, conducted A/B tests and performed statistical analysis of results for multiple features release.
• Performed customer retention analysis resulting in customer segments and identification of the critical factors affecting retention.
• Constructed ETL & BI dashboards on Microstrategy, owned and kept track of 30+ KPIs across pricing, matching, customer experience & safety, analyzed trends, performed root cause analysis and presented reports to the leadership.
• Technical skills: ​Python,​ ​Hive, Qubole, Microstrategy, A/B Tests, Confidence intervals, Clustering, Visualization, Dashboarding

Jun 2016 – Apr 2017 · 10 mos · Bengaluru Area, India

• Developed a merchant churn prediction model( 84% accuracy) using a random forest classifier with an RoI of $20k every month.
• Detected fraud patterns using anomaly detection and identified the areas of potential fraud preventing more than $30k in losses.
• Analysed search engine statistics leading to the introduction of a new feature on the home screen which improved click-through rate by 4%.
• Built rating algorithm for drivers based on their interaction with the system & customer feedbacks- taken into account for their payout.
• Helped the company successfully meet growth objectives through data-driven strategies leading to acquisition by an industry leader.
• Wrote Extract Transform Load (ETL) scripts, analyzed data, drew insights & developed recommendations for business decisions.
• Developed performance metrics dashboards & visualizations on Periscope and presented weekly reports to the leadership.
• Technical skills: ​Python, Pandas,​ ​Numpy, Sql, R,​ ​Dplyr, ggplot, Periscope, Confluence, MS Office