FULL-TIME / CONSUMER FINTECH / 50M+ MAU

Bajaj Finserv: Search Rebuild and One Release Train. Two app stores, two release trains, and 50M+ users waiting on every fix.

+43%Search cross-sell revenue

Bajaj Finserv / Search Rebuild and One Release Train

Role
Software Engineer (Jul 2022 to Jul 2024), then Technical Program Manager (Jul 2024 to Apr 2025)
Timeline
Jul 2022 to Apr 2025, Pune, India
Stack and methods
Angular PWA, Adobe AEM, Adobe Target, ML-ranked retrieval, GPT API
  • +43% search cross-sell revenue
  • ~80% fewer store releases
  • -27% P1 recurrence
  • +64% session depth

Two trains, one release

Interactive / try it

As Technical Program Manager (Jul 2024 to Apr 2025)

  • Owned delivery of Search, Home page, Home Loan, and an AI chatbot across Android, iOS, and Web with 2 PMs, 3 designers, and a 6-engineer pod.
  • Search-origin cross-sell revenue +43% by integrating ad and cross-sell placements into a rebuilt Search surface.
  • Rendering speed +23%, correlated with +64% session depth (tracked via LCP, FCP, crash-free sessions).
  • Consolidated separate Android and iOS release trains into one Angular PWA, cutting app store releases ~80%.
  • P1 incident recurrence -27% via a weekly reliability review, after authoring 6 post-mortems.
  • Release slippage -30% via a cross-team dependency framework that surfaced critical-path risks early.

As Software Engineer (Jul 2022 to Jul 2024)

  • View-cache and async-API library adopted by 25+ teams, cutting redundant API calls 35% and keeping core flows usable on low-connectivity networks.
  • Templated Adobe AEM category landing pages scaled to 12 product lines from one codebase, no per-product build.
  • 10+ features for 50M+ users, including a Search overhaul to ML-ranked retrieval and a GPT-API chatbot.
  • Loans funnel completion +18% via Adobe Target A/B tests.

Decisions worth asking about

Decision 01

One PWA instead of two native release trains.

Store releases dropped ~80% and fixes moved off store review onto direct code push.

Decision 02

Make reliability a weekly ritual.

After 6 post-mortems, a weekly reliability review cut P1 recurrence 27%.

Decision 03

Monetize the surface people already use.

Ad and cross-sell placements inside a rebuilt Search lifted search-origin cross-sell revenue 43%.

← All workRun Fit Check