Gautam Singhania

Product Clustering: Created a Data Science algorithm to cluster similar products across Tesco catalogue for >150K products. Helped Tesco implement fair Price Policies across stores and uphold customer trust. Used a mix of Image Embedding similarity and hard-coded rules with Unsupervised Learning. Achieved an accuracy of 55% direct approval rate and 35% approval rate with amendments, with only 10% rejection rate. Attribute Correction: Worked on various algorithms using Python to enrich product attributes in Tesco catalogue for >500K products. Handled important fields like commodity codes, brand, GPC bricks, and packaging attributes. Helped Tesco avoid legal penalties and smoothens business functions. Used a mix of Image similarity, Text similarity, hard-coded rules, Self Supervised Learning and Supervised Learning. Data Enrichment Platform: Constructed a end-to-end platform to handle Enrichment of data, adding missing data and correcting existing data. Deployed on Azure Kubernetes with Docker, Jenkins, Couchbase and PostgreSQL (SQL). Managed data sync from all sources and orchestrated AI algorithms for enrichments. Provided a place for peer colleagues with business expertise to give feedback on algorithm outputs. Supplier Queries: Maintained an online web portal for suppliers to raise their grievances and get quick support. Introduced workflows like document retrieval and FAQs, automating query resolution by 40%. Showcased a POC on free-text chatbot with RASA that understands intent by training on common supplier queries.

Led the efforts to integrate Android EMM in the existing Android offering by MaaS360.