Headline

MTS-1 @ Nielsen | BITS Pilani ’26

About

I am an avid learner and backend developer focused on building scalable AI applications, currently beginning my next chapter as an MTS-1 at Nielsen. With a dual degree in Mathematics and EEE from BITS Pilani, I enjoy tackling technical constraints to turn concepts into real-world products. Nielsen (Current): Joined as a Member of Technical Staff (MTS-1). Previously, I debugged REST APIs serving 150M+ records, built comprehensive CI pipelines from scratch, and implemented a Gemini-powered code review system. Borderplus: Helped scale a real-time voice-interview pipeline to ~100 daily sessions at <$0.30/session, shipped a low-latency Socket.IO broadcast layer, and built a full-stack platform to define LLM evaluation metrics.

Location

Faridabad

AI stack

antigravity

codex

claudecode

Work experience

Member of Technical Staff-1 at Nielsen, Jul 2026 - Present · 2 mos, Gurugram

Software Engineer Intern at BorderPlus, Jan 2026 - Jun 2026 · 6 mos, Bengaluru, Voice Interview Pipeline at Production Scale Built the core real-time voice pipeline powering the HR-tech product: Azure STT/TTS, Silero VAD for speaker detection, GPT-5 mini as the reasoning layer. Scaled from 6 → ~100 interviews/day at <$0.30/session - the unit economics that made automated interviewing viable. Observer Broadcast System (end-to-end) Designed and shipped the observer architecture: Socket.IO layer streaming Opus-encoded audio to live observers with zero added latency on the candidate session. Enables real-time remote monitoring without touching the interview path. LLM Eval & LLMOps Pipeline Built an evaluation workflow with prompt versioning via Claude Skills (activatable definitions that trigger on prompt changes), plus a pi mono-agent pipeline with read access to production DB that synthesizes eval sets from real data. Nothing untested reaches production. Annotation Platform (Full-Stack) Built FastAPI + React/Vite platform end-to-end. 50 nurses generated 1,500 clinical annotations and 10,000 QnA messages - the dataset now grounding our LLM failure metrics. Granular RBAC, Beacon API for accurate time tracking, referral attribution dashboards. Stack: Python, FastAPI, WebRTC, Socket.IO, Qdrant, PostgreSQL, Azure SDK, OpenAI SDK, Claude SDK, React, Vite

Software Development Intern at Nielsen, Jul 2025 - Dec 2025 · 6 mos, Gurugram, API Debugging at Scale Diagnosed and fixed 4 REST APIs serving 150M+ record datasets to 90% unit-test coverage. Root-caused broken pagination and missing datetime filters that were forcing downstream consumers into full-table scans. CI Pipeline from Scratch Built end-to-end CI covering code quality (gofmt, gocyclo), security (Trivy, Gitleaks, gosec), coverage velocity, mutation testing (Gremlins), and Docker image scanning. Standardized for the team. AI Code Review (adopted by team) Shipped a Gemini-powered code review system that cut review cycles from 2 hours to minutes on clean repos eliminated the pattern of reviews being skipped entirely under deadline pressure. Stack: Go, Python, Docker, Trivy, Gitleaks, Gremlins, Gemini SDK, WhisperAI

Education

Bachelor of Engineering, Electrical and Electronics Engineering, Birla Institute of Technology and Science, Pilani

Master of Science, Mathematics, Birla Institute of Technology and Science, Pilani

Public links

github: https://github.com/Nikesh-1904