Two years building at NielsenIQ

I developed features for the Discover application, and worked on proofs of concept and R&D, including integrating LLMs into Discover. One engine I built cut ticket resolution from 2 days to real time.

Software Engineer, Platform API, NielsenIQ2023–2025

Diagram comparing manual SQL taking 2 days with the recursive engine resolving nested hierarchies in real time
Diagram comparing manual SQL taking 2 days with the recursive engine resolving nested hierarchies in real time.

From July 2023 to July 2025 I worked as a software engineer on NielsenIQ’s Platform API team in Chennai. My work had two sides: building features for a product, and exploring what could come next.

Features for Discover

I developed features for the Discover application, including:

  • Recursive Dynamic Chars Engine. Backend logic in the Dimensional Metadata Service and Data Factory that resolves nested product hierarchies on its own. It replaced manual SQL work and cut ticket resolution from 2 days to real time.
  • Partial caching. A caching layer based on payload fragmentation and hashing that cut data retrieval latency by 40%.
  • Testing. With other engineers, I extended the JUnit suites for complex backend logic and raised coverage from 60% to 86%.
  • Infrastructure automation. I co-developed a maintenance scheduler that triggers jobs from live container usage.

R&D and proofs of concept

I also worked on research and proofs of concept, delivered in 2-week sprints from idea to stakeholder demo:

  • LLMs in Discover. I worked on integrating LLMs into the Discover application.
  • Smart Select NLU engine. A Python and TensorFlow engine that maps natural-language queries onto complex data hierarchies, with context-aware handling of ambiguous queries.

Recognition

1st prize and the Innovation Award at the NielsenIQ Global Hackfest in 2024, and the Rising Star Impact Award the same year.

More work