Aug 2021 – Jul 2023 · 1 yr 11 mos · Bengaluru, Karnataka, India

• Projects I worked on:
• A system to predict the duration and discount to be offered for different sessions of the clearance sale to maximise the revenue. Implemented online learning for regular recalibration of markdown strategy. Got Walmart Impact Award for this.
• A system to predict the space to be allocated to different items in a retail store in order to maximise the revenue.

Dec 2018 – Jul 2021 · 2 yrs 7 mos

• Led a team of 5 to work on the following projects:
• Designed an algorithm on top of word2vec and CNN to predict whether a given address is a residential location or a commercial one.
• Built a system to divide a given set of shipments into a given number of routes. System would learn the affinity and ground constraints from the historic data and use them while minimizing the total distance travelled in the routes.
• Built a system to map the shipments to an appropriate dispatch center under the ground constraints which are learnt from the historic data.

Apr 2017 – Dec 2018 · 1 yr 8 mos

• Key Projects:
• Built a system to predict the probability that a customer is going to return the shipment being delivered to him. Modelled a classifier on top of spark.
• Designed an algorithm to build singular & non-overlapping pin-code polygons from the noisy coordinates of the shipments belonging to the pin-codes in India
• Built a system to predict whether 2 addresses represent the same geographic location
• Implemented a graphical model to predict the geocode for an address given the historic data of addresses and corresponding coordinates.

Mar 2016 – Apr 2017 · 1 yr 1 mo

• Worked on Machine Learning problems around text
• Implemented CNN based algorithm using tensorflow to predict category and subcategory of a product given the title of the product
• Built a system to predict locality hierarchy given an address. Designed a graphical model to learn the possible localities and their hierarchies from the addresses in a semi-supervised fashion.
• Developed a spelling corrector on the basis of similarity in the pronunciation of the incorrect and correct spelling

Jul 2015 – Mar 2016 · 8 mos · Mumbai Area, India

• Solved problems involving extraction of intelligent information from data and using it to improve the user experience on the website. My work involved machine learning, algorithm development and software development.

May 2014 – Aug 2014 · 3 mos · Sydney, Australia

• ● Researched on clustering non-linear manifolds of different dimensions in huge high dimensional data
• ● Formulated Mathematical model based on Maximum Likelihood Estimation of correlation dimension
• ● Devised hierarchical clustering and normalized cut based techniques for manifold clustering
• ● Technique used and appreciated by biologists and ecologists in Australia and Japan for data visualization

May 2013 – Jul 2013 · 2 mos · Bengaluru Area, India

• ● Detected cancer dots from image of cervix using Harris Corner detection and Morphological operations
• ● Enhanced prediction accuracy using Boosted Random Forest based on CoHOG feature of image
• ● Technique incorporated for Cervical Cancer Detection System at Philips Research, Bangalore