Shishir Kumar Prasad Email & Phone Number

San Ramon, CA, US

San Francisco, CA, US

My team, Content AI, leads various content and recommendation initiatives powered by both classical machine learning and cutting-edge generative AI. Our work encompasses a broad spectrum of ML challenges, utilizing both traditional and state-of-the-art techniques that profoundly impact every facet of grocery e-commerce.Key areas of focus include:*.

San Francisco, CA, US

In my role as Engineering Manager for the Content & Recommendations AI team at Instacart, I lead a high-performing group pioneering advancements in diverse areas such as personalization, recommendations, catalog optimization, and conversational bots. Our team spearheads major initiatives in the fields of large language models, generative AI, and.

San Francisco, CA, US

Tech lead of personalization, recommendations, and quality ML team at Instacart. My role is to drive the technical roadmap and vision for the team, mentor junior engineers, and help grow the algorithms org.Some of the impactful projects from my team include: Quality - Item availability prediction, Replacement recommendationsPersonalization - Multi-stage.

San Francisco, CA, US

Leading full-stack machine learning projects for product discovery and recommendations.

San Francisco, CA, US

* Architected the replacement recommendation ML system. This is a critical system that ensures we surface the most relevant substitute recommendations when grocery items are out-of-stock in the stores.

San Francisco, CA, US

Building data products at the intersection of machine learning and distributed systems.

San Francisco, CA, US

• Working on the real-time ingestion and catalog data infrastructure team.
• Built a tool to detect near-duplicate images in catalog using fingerprinting techniques like Phash.
• Built predictive models for supply chain inventory management using state-of-the art machine learning systems like Vowpal Wabbit, XGBoost etc.

Bentonville, Arkansas, US

• Architected, Designed and Implemented the business-critical product matching platform that powers the duplicate product detection functionality on Walmart.com.
• Built a real-time big data pipeline to find similar products with high accuracy.
• Used state-of-the-art duplicate detection algorithms like LSH, Minhashing, Image Hashing etc.
• Learnt supervised machine learning models to predict similar products with high precision.
• Integrated with crowdsourcing systems to measure prediction accuracy and other metrics.
• Maintained the Elasticsearch cluster, used for finding candidate products for matching.

Madison, WI, US

• Worked with Prof. Anhai Doan on auto-generation of entity matching rules using supervised machine learning models.
• Designed an entity matching rule management platform.Technologies involved: Java, Weka, Play framework, Twitter bootstrap, MySQL.

Bentonville, Arkansas, US

• Designed a Domain Specific Language (DSL) to write rules for product matching. Used Javacc parser generator to implement the compiler and grammar.
• Developed a web application to help analysts write business rules using above DSL. Added features to recommend rules for matching products based on training data and quickly tune the rule parameters.Technologies.

New York, NY, US

We developed an ETL (Extract, Transform and Load) application called SEED to collect all the transactional data of the firm, processed it and persisted it in a central data warehouse that served enriched data to all the in-house applications across the firm.

New York, NY, US

We developed the reporting infrastructure and catered to business intelligence requirements of the firm like generation of reports, interactive querying, ETL solutions using SSIS workflows and OLAP multi-dimensional cubes to access aggregates information across millions of records.

Masters Of Science(Ms), Computer And Information Sciences

Bachelor Of Technology ( B.Tech.), Computer Science And Engineering

Aissce, Non-Medical Stream