• Contributed to core infrastructure for KYC Refresh platformization across 22 repos, authoring 200+ PRs, including query services, cron triggers, and geo-specific strategy patterns for 6 international regions

• Delivered query optimizations in a year long initiative by analyzing 2.5k+ SQL queries, reducing daily refreshes from 8k to 2k and unrestricting 3.2M+ users.

• Built multi-channel notification system (email, push, in-app) with notification service integration, coordinating with Legal, Design, and Product to ship 45+ templates with feature-flagged rollout

• Partnered on DynamoDB migration, owning implementation of Optimistic Concurrency Control (OCC) to prevent lost updates, and adding integration tests for race conditions and version validation

• Handled 50+ days of on-call coverage, owning resolution of a Sev5 incident and also built a lightweight on-call automation tool using Datadog MCP and LLMs to reduce alert fatigue by summarizing monitors and suggesting initial RCA.

• Secured third place in an on-chain security-focused CTF competition winning 0.25 ETH.

• Worked with Coinbase Wallet team, developing features to onboard crypto into user wallets and enable cashout of crypto to fiat

• Launched cbBTC cashout functionality in the wallet app, driving ~$1M in transactions during the US elections week.

• Contributed backend support for gas-fee-sponsored cashouts, now 15% of wallet app transactions.

• Resolved critical bugs to enable seamless USDC-to-Instant ACH and ERC20-to-ACH cashouts.

• Worked on a visual builder tool aimed at enhancing content creation and website customization

• Built an end-to-end timeline feature allowing users to preview upcoming changes

• Extended invoice charge functionality by introducing optional parameters to give customer more flexibility while collaborating with cross functional teams

• Ensured correctness by implementing unit tests with coverage increase from 38 to 73 percentage

• Developed MVP of a new product aimed at optimizing daybook keeping service, by automating entry of key-value pairs from invoices with OCR from text extraction and GPT for processing text

• Implemented solution made use of asynchronous processing of requests which reduced processing time by 10 times and was 5 times cost effective than industry solutions like Nanonets