my career has moved across security research, product engineering, customer architecture, independent building, and applied ai. the path has not been linear, but the transitions have generally followed the same pattern: one experience exposed a new class of problems, and i moved closer to the part of the work where i believed i could contribute most.
i currently work across several product and platform areas at meshy.
i helped take hibe, an ai application builder, from a rough prototype toward external usability. i also built agentfn, a python framework for exposing coding agents such as claude code and codex as typed function calls for reusable application workflows.
my other work includes developing a cohesive content-moderation and policy-enforcement system, along with an agentic advertising-workflow platform for internal growth users.
during this period i built and launched ai products and technical experiments across 3d generation, document workflows, education, video, voice, and data tooling.
blocksmith became the deepest of these projects. i started with no 3d background, built and shared an early prototype, changed the underlying representation based on developer feedback, and eventually shipped a full product used by game-development teams.
i later sunset the commercial product and open-sourced its core after concluding that the market and investment required for the next quality leap did not justify continuing it as a company.
this period also produced omniflow and omnilaunch, which grew out of recurring problems i encountered while preparing multimodal data and running model experiments. more of the apps and prototypes are on the projects page.
eyris was exploring enterprise-security applications for private blockchain networks. i deployed and automated a fault-tolerant besu/ibft network, then worked backward from customer questions to build demos around secure exchange, auditing, and privacy-preserving communication.
one experiment used zero-knowledge proofs to explore communication between medical providers and insurers without exposing the underlying sensitive data.
i completed the six-month engagement and returned to independent product work.
stripe was my first role centered on customer discovery rather than primarily implementation. i joined pre-sales calls as the technical expert, learned how prospective customers handled payments, and mapped those workflows onto stripe’s product and api surface.
the work involved identifying when product, legal, or other specialists needed to be involved, building demos, and turning ambiguous requirements into a concrete integration plan that a customer’s engineering team could execute.
the role strengthened my ability to distinguish requested features from underlying workflow problems, but i eventually realized i wanted to move closer to implementation again.
towards the end of my time there, i - with my manager’s support - began talking with engineering teams about moving back into a software-engineering role. i found multiple potential teams, but company-wide hiring and transfer freezes prevented the move. stripe later announced its november 2022 layoffs, which affected me and most of my team.
tabu art was my first full end-to-end production product and startup. i taught myself django, aws, ethereum development, payments, protected media delivery, and android development while building the marketplace.
the product included credit-card checkout over blockchain settlement, royalty routing, buyer-funded minting, protected high-resolution assets, and an android tv experience.
we built too much before validating the assumptions that mattered most, and i eventually stepped away after our product and financing views diverged. i transitioned the system to another engineer before leaving. there is more detail on the projects page.
raytheon was my first professional engineering environment. after teaching myself programming, linux, exploit development, and computer-security fundamentals in college, i joined a small team focused on vulnerability research.
the work became increasingly open-ended and gave me a deep grounding in c, assembly, operating systems, networking, embedded environments, low-level debugging, reverse engineering, fuzzing, and custom research tooling. i later moved into development work spanning embedded realtime systems, mobile devices, and higher-level applications.
i eventually got the chance to lead a high-impact project during a period of serious delivery risk. the team had been working on the product for years and was incredibly talented, but the engineers were dealing with organizational problems that made it difficult to focus on delivery. my job was to get the project back on the rails and remove obstacles so they could do their best work.
it was tough taking a step back from technical work, but highly rewarding to help turn things around. we delivered more than five major releases, maintained a strong customer relationship, and retained a key engineer for the project.
before leaving, i found someone to shadow me and handed the role off gradually so the team would not face another leadership gap.