Deep dives
Problem, context, constraints, architecture, decisions, trade-offs, and outcome across professional and personal engineering work.
Personal project: treating retrieval infrastructure as a lifecycle problem, not a demo
A self-initiated, local-first RAG backend built to explore what production-grade retrieval infrastructure actually requires beyond a working demo: lifecycle management, multilingual retrieval, and framework independence.
Personal project: treating AI agents as an SDLC, not a code-completion tool
A self-initiated project exploring what happens when a requirement flows through a pipeline of specialized AI agents (planning, implementation, testing, review) end to end, instead of a single assistant writing code from a prompt.
Delivering a defined multi-boundary configuration flow reliably, not redesigning it
Building a React/Node configuration dashboard that lets application settings be updated live in Kubernetes, by loading a secret from HashiCorp Vault, running an AWS Lambda, updating a Kubernetes ConfigMap, and rerunning the pipeline.
Validation, defensive state handling, retry-eligibility semantics, and scheduled cleanup in a live Java payments platform, plus cross-service payment-state debugging
Hardening validation and correctness, owning the backend retry-eligibility and failure-taxonomy contract for failed credit-card payments, and automating lifecycle cleanup in Intuit's QuickBooks Bill Pay B2B payments service: batch-validation and payment-state-aware abort fixes, idempotent downstream QuickBooks booking that keeps working after execution and bookkeeping lose their shared event, a canonical failure-to-retryability mapping exposed as typed failure metadata on the batch-payment polling API, a runtime-refreshable Payment Instruction expiration TTL, a ShedLock-coordinated scheduled stale-Payment-Instruction cleanup job built on a self-designed sparse, sharded DynamoDB GSI with authoritative-state revalidation before every abort, a failed-credit-card read-path filter, and the end-to-end investigation and remediation of a cross-service payment-state defect, where inconsistent payment-state interpretation across asynchronous paths produced incorrect notification behavior for mixed-outcome batches, backed by regression and production-like end-to-end validation.
Primary-engineer ownership of a shadow-validated, agent-assisted cutover, not a rewrite
Leading the migration of a set of Tier-1 financial-institution data-provider integrations onto Intuit's modern execution engine as primary engineer: a Claude-powered migration agent for combined banking-and-tax providers, hand migration for tax-only scripts, and shadow validation against each provider's legacy path before any production cutover.
Independent architecture ownership, validated after the fact
Designing and owning a centralized Kafka routing layer for a banking platform, independently, and having the design validated by the principal architect after the fact rather than before.