Led a full-stack team shipping consumer apps on top of Vana’s protocol. I focused on translating Web3 primitives into UX that felt normal to Web2 users, without compromising on privacy or permissions. Owned the Vana App and staking products, scaled the applications layer to 100k+ users, and set up PostHog and Sentry so we caught issues before users did.

Later took over as technical lead and ran the overhaul of Vana’s Data Portability protocol: wrote the spec, got the team on the same page, and shipped the implementation. Owned delivery across TypeScript and Python services, Postgres, and the infra surface (GCP + Cloudflare). Sped up PR throughput and time-to-prototype by pushing clearer specs and tighter review loops. Wrote the kind of end-to-end tests and docs that engineers actually use.

Joined as an early engineering hire. Designed a patented method for encrypting large datasets using a Web3 wallet (non-custodial encryption and decryption), then built the data wallet and secure APIs around it. Also built a framework for offline, in-browser data processing so we didn’t have to route everything through centralized services. Audited cloud spend and cut the AWS bill by roughly 90%.

As technical lead, took an AI image generation app from idea to launch in two weeks, then kept it running as usage grew past 1M users. Rebuilt model-wrapping APIs as models improved, cutting waste in storage and compute. Wrote design docs that let frontend and backend work in parallel instead of handing off sequentially. Ran technical interviews and helped grow the engineering team.

Moved into engineering management, leading a team across backend, frontend, data, and ML. We built an API for third-party developers to deliver personalized AI experiences, with privacy and security as hard requirements. That included an OAuth-based permissioning system for sensitive user data. Wrote and maintained API docs so integration didn’t require tribal knowledge. Worked with leadership to keep platform work pointed at actual user problems.

Built an MLOps pipeline from scratch (Terraform, GitHub Actions, SageMaker, Python) so ML models could ship like real software. Wrote production Python APIs around those models with clean contracts for data science teams and downstream consumers. Shipped features that cut manual work for users: document verification, passport OCR, plagiarism checks.

Fixed performance problems (N+1 queries, slow endpoints under load) and worked across Python, Ruby, TypeScript, React, Docker, and Kubernetes depending on what the problem needed. Also ran system design interviews.

Started out building internal and consumer-facing systems for a major automaker using Spring (Boot, MVC, Data, Security), MS SQL, and AWS. Shipped Spring microservices in Docker with CI/CD in Azure DevOps. Pushed test coverage and code review standards up, which cut regressions and made releases less painful. Mentored newer engineers and helped establish reusable patterns across projects.

Moved into mobile, shipping Android and iOS apps for enterprise clients and startups with Xamarin, REST, and JWT. Ran bi-weekly demos with clients and iterated fast, which built trust and led to bigger engagements. Tightened up app security (session timeouts, secure storage, safer caching) and guided clients through App Store submissions so launches went smoothly.

Eventually took on a management role. Led multiple client deliveries while still writing code in Spring Boot, Angular, Node, Salesforce, and Mulesoft when needed. Set up CI/CD pipelines in Azure DevOps and Terraform-based infrastructure-as-code so environments were reproducible. Worked with AWS architects on system reviews, and led the effort to get AppCentrica its AWS partner designations. Ran hiring, onboarding, 1:1s, and performance reviews.

Built and maintained an automated test framework for installing and validating WebSphere Application Server across multiple platforms. Added installation and functionality test scenarios, with a focus on the edge cases that usually slip through. Wrote Windows and Unix scripts to automate repetitive team tasks and make test runs more reliable. When automation broke, triaged the failures and either fixed the framework or filed clear repro steps against the product.

Came back as a software engineer working on install, upgrade, and maintenance flows for WebSphere (Network Deployment and Liberty). Spent most of my time in IBM Installation Manager, debugging real-world failures: upgrades that broke, rollbacks, config drift, platform differences. Cut install time by trimming unnecessary steps in the install path. Wrote build and release automation in Java, Ant, Maven, and some JavaScript tooling to make releases less manual.