Avish Bhasin
With more than 5 years in data science and analytics, I specialize in transforming complex data into actionable strategies. My journey spans high-impact roles at Uber and Blinkit, where I've harnessed ML and AI to optimize operations, enhance user experiences, and fuel business growth. From building demand forecasting systems to optimizing delivery logistics, I bring a hands-on approach to solving real-world challenges. Whether you're aiming to elevate your analytics skills, explore career pathways in data science, or learn how to turn data into strategic advantages, I’m here to mentor you on your journey to success.
REVIEWS FROM CLIENTS
Built a propensity model using XGBoost to target new users, driving initiatives like promos, cancellation waivers, and surge waivers, ...
Built a propensity model using XGBoost to target new users, driving initiatives like promos, cancellation waivers, and surge waivers, which resulted
in a 5% increase in trips, 8% growth in Gross Booking V alue ($30 million), and a 2.5x return on investment.
● Optimised ML-Driven Demand Forecasting: Developed an advanced forecasting system for 304 stores, integrating Facebook’s Prophet and ...
● Optimised ML-Driven Demand Forecasting: Developed an advanced forecasting system for 304 stores, integrating Facebook’s Prophet and Google's
Temporal Fusion Transformers to enhance forecast accuracy by up to 25%. Utilised LightGBM (LGBM) models to improve prediction precision for
low-velocity items. Achieved substantial cost savings of ~2 crore annually while elevating product availability to 95-98%, directly impacting
profitability and customer satisfaction.
● Identified and clustered synonymous search terms in blinkit search queries using NLP techniques (NLTK, Gensim, and Transformers), reducing
null/empty searches by 30% and thus enhancing product discoverability and inventory assortment. Achieved a 7% increase in platform availability
and revenue boost of INR 10 lakhs monthly through improved assortment planning.
● Blinkit Delivery Bags Optimisation: Developed a 3D bin packing model to optimise delivery bag usage, achieving space utilisation efficiency and cost
reduction using techniques like dynamic programming and simulations. Resulted in a significant reduction in bag costs by 1 crore monthly. Utilised
the model to create efficient gifting bag options, enhancing the customer experience on the app.
● V ehicle Routing Algorithm: Created an OR based VRP model for the cold and frozen supply chain which led to decrease in the vehicles deployed
(~40%) and hence a cost saving of approximately ~25 Lakhs monthly.
Network Planning algorithm: Worked on an optimisation algorithm for Darkstore - Warehouse mapping based on distance and feasibility us...
Network Planning algorithm: Worked on an optimisation algorithm for Darkstore - Warehouse mapping based on distance and feasibility using linear
programming techniques. The algorithm aided in strategic decisions for opening/closing backends in various locations, saving ~5 Lakhs monthly.
● Middle mile Vehicle planning: Refined vehicle demand creation logic, leading to ~ 4 Lakhs per month in cost reductions and improved vehicle
utilisation. Also, developed a CatBoost model for predicting ETA of fleet vehicles with ~ 83% accuracy.
● Managed analytics for both operational and product areas in the middle mile team. Key contributions included scaling the app, implementing cost
management, improving crate utilisation, and developing new product constructs to boost operational efficiency and innovation.
● NU Delivery and Surge Charge Removal: Designed and implemented the experimentation of the removal of surge and delivery charge for new users by
using A/B testing and causal inference techniques. This helped in the growth of top of the funnel metrics with an increase of ~5% in conversion levels
for the target group.