Headline

Senior SDET @ CRED | 0-to-1 rewards & gamification since launch, IPL campaigns, Credit card bill payments | Built CRED’s AI test automation tooling and an autonomous release agent(~15h saved/release)

About

Eleven years in engineering. I've built the test automation framework at every company I've worked at — Selenium for IHG at Accenture, JavaScript/Cypress at Bank of America, and three generations of mobile automation at CRED. Each time the work moved up a level — writing the tests, owning the framework, rebuilding it with AI, then building the agents that operate it. Five years at CRED. First on Win — CRED's rewards and gamification surface — from its launch in 2021 through 2024: Spin the Wheel, Arcade, Mystery Reward, Jackpot, Scan & Pay, the redemption flow, the Voucher and Wallet Farm revamps, plus multiple GTM launches and successive years of IPL campaigns. Since late 2024, Card Management and Credit Card Bill Payments. Both charters ship to CRED's full member base, on surfaces that move real money. Gen 3 of the automation is an AI-driven Python platform . I improved the AI locator engine that resolves elements in three tiers (primary → fallback → vision), contributing the prompt system for locator accuracy, element-type weighting for candidate ranking, and an intelligent retry engine. The piece I'm proudest of is the eval harness — golden dataset, CLI with self-test and accuracy gating, documented metrics. It's what makes "AI-powered testing" measurable instead of asserted. Building an AI feature is easy now. Proving it works isn't. Then the agent layer. I own an autonomous agent in production that orchestrates iOS and Android releases — it tracks 22 LOB sign-offs, triggers builds and monitors readiness, replacing roughly 15 hours of manual coordination per release. Earning the right to act autonomously was the hard part: shadow mode, progressive canary unlock, separate write and outbound gates, identity guardrails. I also own an agent that surfaces untested critical paths in pull requests, and I've built six MCP servers and four skill packages on CRED's shared agent platform. One technical opinion I've held for four years: fix test determinism at the data layer, not with retries at the assertion layer. Certified: Upraised Product Management (2024) Happy to talk about consumer fintech, giving autonomous agents real permissions safely, and what "launch ready" should actually mean.

Location

Bangalore Urban

AI stack

anthropic apis

claude code

claude

github copilot

n8n

obsidian

Work experience

Senior SDET-Card Management & Credit Card Bill Payments at CRED, Sep 2024 - Present · 2 yrs, • Built the eval harness that makes AI-driven testing measurable — golden dataset, CLI with self-test and accuracy gating, documented metrics. • Built and own an autonomous release orchestration agent used by mobile release managers — tracks 22 LOB sign-offs, triggers builds, monitors readiness, replacing ~15 hours of manual coordination per release. Designed its progressive-autonomy rollout: shadow mode before it could act, canary levels unlocking capability in stages, separate gates on database writes and outbound actions, and identity guardrails preventing dispatch of the wrong build. • Contributors to CRED's codeless mobile automation platform — improved the AI locator engine (prompt system, element weighting, retry engine), shipped across its three execution tiers, and authored the modernization roadmap that moved scheduled runs in-house. • Built and run the release process for Card Management — Jira dashboard with daily P0/P1 visibility, sign-offs gated before every release, sprint backlog grooming, and the test estimation template adopted across two PODs. • Own quality and release readiness for CCBP and Card Management, shipping to CRED's full member base — 18+ features including the Payment Bottomsheet Flutter revamp, CCBP Autopay, BBPS bill fetch, RuPay and CP silent linking, Smart Statement V2, Cards Revamp 4.0, and AMEX and Visa in-app flows. • Built the tooling the team runs on — WireMock stubs for issuer-failure simulation, a single source of truth for test data, an FTUE checklist, and data-seeding utilities other engineers' tests depend on across a 10+ person repo. Own automation and CRT regression cycles. • Built an agent surfacing untested critical paths in PRs with ROI tracking, six MCP servers and four skill packages. Across five years: 300 commits over four automation platforms in Python, Java, TypeScript and YAML; 660+ defects driven to closure, 1000+ test cases; interns and junior engineers mentored.

Senior SDET-Win (Rewards & Gamification) at CRED, Oct 2021 - Aug 2024 · 2 yrs 11 mos, Bengaluru, • Quality owner on Win, CRED's rewards and gamification surface, from its launch through three years of iteration — shipped 0-to-1 mechanics including Spin the Wheel, Arcade, Mystery Reward, Jackpot, Scan & Pay, the redemption flow, and the Voucher Farm and Wallet Farm revamps. • Shipped 17+ features end-to-end across Android and iOS — Pop Challenge V2 plus its explainer and in-app redemption, CLAW, the AR Game, Card Linked Offers, cashback enhancements, PPS metrics and voucher changes, and voucher redemption with claim-rate and bump-rate tracking. • Quality owner for multiple GTM launches and successive years of IPL campaigns — CRED's highest-traffic, hardest-deadline moments, where the window to catch a defect closes before the campaign does. Worked across Firebase, Mixpanel, Datadog RUM, Lottie, Jetpack Compose, UPI SDK and in-app webview surfaces. • Laid the automation foundations that became the 2026 platform: battery and performance measurement under varying network conditions, crash-log capture turning failures into diagnosable artifacts, the first image-comparison proof of concept, and the first retry handling for flaky assertions.

Quality Specialist II B at Bank of America, Aug 2021 - Oct 2021 · 3 mos

Senior Software Engineer at Bank of America, Aug 2019 - Aug 2021 · 2 yrs 1 mo, Gurugram, • Built and maintained the JavaScript + Cypress end-to-end automation framework — authored and extended test suites, kept them stable as the application changed, and grew coverage across releases. • Owned framework maintenance as well as authorship — the same build-it-then-own-it pattern I've carried through three further automation platforms since.

Analyst at Accenture, Jan 2016 - Jul 2019 · 3 yrs 7 mos, • Built and maintained the Selenium automation framework for IHG (InterContinental Hotels Group) — hospitality booking platform. • Owned the suite across three and a half years: authored coverage, kept it stable through application change, and maintained it as the primary regression safety net.

Software Engineer at QA InfoTech - Your Software Testing Partner, Apr 2015 - Dec 2015 · 9 mos, • Localisation and globalisation testing for Adobe Photoshop — verified UI, layout and content behaviour across locales and regional configurations. • Accessibility testing across the product surface.

Education

Bachelor of Technology (B.Tech.), Electrical, Electronics and Communications Engineering, Ims Engineering college Ghaziabad

Sunbeam English School Bhagwanpur Varanasi