Vandana Kumari profile photo 1 on Tal

Vandana Kumari

Software Engineer, PdE @Mercari | Building Global(Cross border) Product at Scale | Java, GoLang, Microservices, Distributed Systems, AI Native | ex - @UniAcco @Wozo Gaming | BTech CSE

Gurugram

3+ years of work experience

Mercari India
Feb 20267 mosPresent

Built an Image Search system enabling users to discover similar products using images by designing scalable search APIs and integrating similarity-based retrieval, embeddings, vector DB across Mercari catalog services, increased ¥77.4M/month GMV lift and 1,801% ROI, used by 2M+ monthly search UUs. Authored TDDs to drive architecture decisions and align with business requirements. • Expanded scope into full-stack product engineering - independently driving frontend development, PRDs, EDDs, and UI/UX design end-to-end using AI-native development workflows (Cursor, Copilot), shipping features E2E at accelerated pace; delivered Like utility features with +30.7% / +29.6% / +40.1% improvement (iOS/Android/Web) independently. • Developed the Global One Product app, owning the backend services. Delivered 6+ features including partial SDUI homescreen, deep link generator, account management, checkout service, item details screen & SERP(SpringBoot). • - Launched in HK, TW and US, cutting per-market localization integration effort by 40%. • Led development of search and metadata APIs and a notification generator service in Java, directly improving product GMV; • Integrated circuit breakers across multiple upstream microservices, tuned error threshold from 1,000 → 5,000, reducing CB-triggered errors by 65% in production, improving system resilience and fault tolerance. • Implemented SLOs, monitors, and PagerDuty alerts via Datadog across Backend, BFF endpoints - deployed error budget burn alerts and reduced unnecessary on-call pages by 50% via threshold tuning, pages, significantly improving observability and incident response time. improved Image Search latency SLO from <1% → 65% compliance (p95 upper bound 1.5s). • Created CI/CD pipelines and automated database schema migrations using Goose for master service via environment-scoped GitHub Actions workflows (dev + prod) reducing manual deployment effort by 60%, enabling safe version transitions.
Mercari India
Jun 20241 yr 9 mosFeb 2026

Mercari India
Jan 20246 mosJun 2024

UniAcco
Jun 20233 mosAug 2023

Mumbai, Maharashtra, India

Protégé IGDTUW
Feb 20235 mosJun 2023

Wozo Gaming
Oct 20223 mosDec 2022

INDIRA GANDHI DELHI TECHNICAL UNIVERSITY FOR WOMEN
Jun 20222 mosJul 2022

New Delhi, Delhi, India

Vandana’s AI Stack

codex
codex
cursor
cursor

Education

Indira Gandhi Delhi Technical University for Women

Bachelor of Technology - BTech, Computer Science

Indira Gandhi Delhi Technical University for Women

20204 yrs2024
kendriya vidyalaya

Xll, PCM with computer science

kendriya vidyalaya

Join Vandana and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Vandana Kumari on Tal

Headline

Software Engineer, PdE @Mercari | Building Global(Cross border) Product at Scale | Java, GoLang, Microservices, Distributed Systems, AI Native | ex - @UniAcco @Wozo Gaming | BTech CSE

About

I build backend systems that scale — and ship features that move numbers. At Mercari, I engineered an Image Search system now driving ¥77.4M GMV/month at 1801% ROI, serving 2M+ users across one of Asia’s largest C2C marketplaces. I’ve contributed to multi-country product launches (HK, TW, US), improved latency SLOs from <1% → 65% compliance, and owned features end-to-end — from system design to production monitoring. Beyond backend, I work across the frontEnd using AI-native workflows (Cursor, Copilot) and have led org-wide AI initiatives as a single-stakeholder pDE — driving POCs that cut across product, design, and engineering. I’m deeply interested in GenAI and LLMs & RAG — not just as a trend, but as tools I actively integrate into how I build and ship products faster. Outside of building: I conduct engineering interviews, mentor junior engineers, and contribute to org-wide growth and planning initiatives. Outside of Work - I love writing my thoughts, Learnings & Sharing what’s beneficial for others that I could know earlier. My Medium Blog - https://medium.com/@vandana.igdtuw/i-built-an-image-search-system-heres-what-i-actually-learned-b56abacdc06f

Location

Gurugram

AI stack

codex

cursor

Work experience

SDE-2(AI Native) at Mercari India, Feb 2026 - Present · 7 mos, Built an Image Search system enabling users to discover similar products using images by designing scalable search APIs and integrating similarity-based retrieval, embeddings, vector DB across Mercari catalog services, increased ¥77.4M/month GMV lift and 1,801% ROI, used by 2M+ monthly search UUs. Authored TDDs to drive architecture decisions and align with business requirements. • Expanded scope into full-stack product engineering - independently driving frontend development, PRDs, EDDs, and UI/UX design end-to-end using AI-native development workflows (Cursor, Copilot), shipping features E2E at accelerated pace; delivered Like utility features with +30.7% / +29.6% / +40.1% improvement (iOS/Android/Web) independently. • Developed the Global One Product app, owning the backend services. Delivered 6+ features including partial SDUI homescreen, deep link generator, account management, checkout service, item details screen & SERP(SpringBoot). • - Launched in HK, TW and US, cutting per-market localization integration effort by 40%. • Led development of search and metadata APIs and a notification generator service in Java, directly improving product GMV; • Integrated circuit breakers across multiple upstream microservices, tuned error threshold from 1,000 → 5,000, reducing CB-triggered errors by 65% in production, improving system resilience and fault tolerance. • Implemented SLOs, monitors, and PagerDuty alerts via Datadog across Backend, BFF endpoints - deployed error budget burn alerts and reduced unnecessary on-call pages by 50% via threshold tuning, pages, significantly improving observability and incident response time. improved Image Search latency SLO from <1% → 65% compliance (p95 upper bound 1.5s). • Created CI/CD pipelines and automated database schema migrations using Goose for master service via environment-scoped GitHub Actions workflows (dev + prod) reducing manual deployment effort by 60%, enabling safe version transitions.

Software Engineer at Mercari India, Jun 2024 - Feb 2026 · 1 yr 9 mos

Software Engineer Intern at Mercari India, Jan 2024 - Jun 2024 · 6 mos

Software Engineer Intern at UniAcco, Jun 2023 - Aug 2023 · 3 mos, Mumbai, Maharashtra, India

Mentor at Protégé IGDTUW, Feb 2023 - Jun 2023 · 5 mos

Software Developer Intern at Wozo Gaming, Oct 2022 - Dec 2022 · 3 mos

Machine Learning Intern at INDIRA GANDHI DELHI TECHNICAL UNIVERSITY FOR WOMEN, Jun 2022 - Jul 2022 · 2 mos, New Delhi, Delhi, India

Education

Bachelor of Technology - BTech, Computer Science, Indira Gandhi Delhi Technical University for Women

Xll, PCM with computer science, kendriya vidyalaya