Engineering Blog
Deep dives into the systems behind shipped products. Monorepo patterns, real-time architecture, browser automation, and what actually works when you put AI in front of users.
May 25, 2026
I started as a third-party integrator before shifting into full-stack platform development. Here is how I decide when to build CRM workflows, when to use Dynamics 365, NetSuite, or Intacct, and how to extend enterprise systems when bespoke needs arise.
May 24, 2026
Our platforms were stable and revenue-positive, but scale exposed retry storms, fanout bottlenecks, and noisy incidents. Here is how we decided when to add BullMQ and RabbitMQ, what we measured, and what changed.
May 23, 2026
Platforms like LibreChat give you a production-ready chat UI, multi-model routing, and built-in Langfuse integration. Custom agent frameworks give you reactive UI, transactional tool execution, and full rendering control. Here is how I decide between them across four different products.
May 22, 2026
Microfrontends are not a default architecture. Here is how I decided to migrate from a frontend monolith to domain-based frontend apps, what changed in deployment speed and cost, and when this approach is worth it.
May 21, 2026
I built an AI evaluation framework from scratch to understand every piece — ground truth datasets, precision/recall harnesses, LLM-as-judge layers, ClickHouse dashboards. Then I migrated it to Langfuse. Here is why, and what I gained.
May 21, 2026
After years running companies, I intentionally moved into a Senior Full Stack Developer role to close technical gaps, learn inside a larger organization, and build truly scalable platforms from the ground up.
May 20, 2026
Part 1 of my portfolio AI series: Convex agent component design, RAG fundamentals, OpenAI vs Google provider testing, streaming UX, and how the frontend chat bubble and /admin/ai tooling live inside one reusable AI plugin.
May 19, 2026
Part 2 of my portfolio AI series: concrete eval results, ClickHouse monitoring signals, optimization decisions, recursion guardrails, and the measured before/after impact.
May 18, 2026
A practical guide to embedding dimensions using our Portal support retrieval pipeline and shared support/AI components: how dimension choice changes quality, latency, storage, and migration strategy.
May 17, 2026
Most developers use AI to generate throwaway code for every new project. I use it to build CLIs, agent skills, test harnesses, and plugin scaffolds — repeatable infrastructure that compounds. Here is why spinning up a fresh agent session per task, with pre-built tools and context, beats ad-hoc prompting every time.
May 16, 2026
Our first AI feature hallucinated a compliance finding that did not exist. A client acted on it. Here is how we built input sanitization, output filtering, confidence scoring, and human-in-the-loop escalation to make AI output trustworthy.
May 15, 2026
Our AI scanner said a page violated WCAG 2.4.7 — but it cited the wrong success criterion. RAG fixed the hallucination problem by grounding every AI claim in the actual specification, with retrieval metrics that prove the system works.
May 14, 2026
AI scanners find accessibility violations that deterministic tools miss — but how do you know the AI is not hallucinating? We built an evaluation harness that measures precision and recall against axe-core ground truth, gates every prompt change with regression tests, and tracks quality drift in ClickHouse.
May 13, 2026
Both Cloudflare Workers and Convex let you run server-side code without managing servers. But they solve fundamentally different problems. After running both in production, here is when I reach for each one — and where they overlap more than you might expect.
May 12, 2026
Monitoring, tracing, and instrumentation look different in every system we run. Here is what we actually measure, how we measure it, and what those measurements have caught — across Kubernetes infrastructure, integration pipelines, high-performance APIs, browser automation, and AI agents.
May 11, 2026
Every system in the LaunchThat ecosystem uses classical software engineering patterns — Strategy, Observer, Adapter, State Machine — and formal data structures. Here is how GoF patterns, SOLID principles, and CS fundamentals show up in production code.
May 10, 2026
Both .NET and Convex can power production backends. After building multi-tenant SaaS platforms, high-performance APIs, and real-time dashboards across both, here is how I decide which one fits — and when the answer is both.
May 9, 2026
We moved off Vercel onto a 128GB bare-metal server. Each client now gets their own frontend containers, their own Convex backend instance, and their own database backups — orchestrated with Kubernetes and Helm. Here is how we got here and what we learned about owning infrastructure.
May 8, 2026
Switching from Drizzle/Postgres to Convex collapsed the frontend-backend gap, doubled development speed, and let me cancel $230/month in third-party services. Here is how the migration worked and what it unlocked.
May 7, 2026
I replaced a WordPress multisite with a custom Next.js app backed by Postgres. Load times dropped from 8 seconds to under 2. But the platform was still fractured across a dozen third-party services, and every schema change was a deployment risk.
May 6, 2026
More than 10 years ago, I built a WordPress field app for a service business while in college. It doubled job throughput, improved margins, and changed my career path from computer engineering to entrepreneurship.
May 5, 2026
Seven years ago I set out to build a multi-tenant platform on WordPress Multisite. All client data lived in one database, third-party glue held everything together, and a $100/mo host still could not keep up. Here is what I learned — and why I had to leave WordPress behind.
May 4, 2026
Neither ClickHouse nor Convex alone solves the trading chart problem. Here is how we split candle data between a columnar time-series engine and a real-time document database — and why the boundary sits exactly where it does.
May 3, 2026
We kept creating new repos for every project. Shared code drifted, configs diverged, and a single bug fix became a multi-day coordination exercise. Here is how a Turborepo monorepo fixed all of it.
May 2, 2026
Our trading dashboard polled every 5 seconds and users complained about stale data. We rebuilt on Convex with real-time subscriptions and the difference was not incremental — it was a different product.
May 1, 2026
Our first AI feature hallucinated a refund policy that did not exist. A customer followed it. Here is what we learned about putting language models in front of real people.
Apr 30, 2026
Every customer wanted something different. The CRM customer needed contacts. The LMS customer needed courses. We were shipping bloated software that served nobody well. Here is how plugins fixed it.
Apr 29, 2026
A customer asked for Monday.com integration. We could have hard-coded it into the core. Instead, we built a plugin system — and 18 months later we have 34 plugins serving 27 apps.
Apr 28, 2026
We embedded a Spline scene in our trading dashboard. The first version ran at 12fps on a MacBook Air. Here is how we shipped 3D that users actually enjoyed instead of endured.
Apr 27, 2026
Our Monday.com plugin made 47 API calls per dashboard load. Users waited 8 seconds for data they had seen 5 minutes ago. Here is how we built a caching layer that made the integration feel instant.
Apr 26, 2026
Monday.com's app framework promises drag-and-drop integration. The reality involves undocumented OAuth quirks, a pagebuilder that fights your layout, and an approval process that rejects you for reasons you cannot reproduce.
Apr 25, 2026
A Puppeteer script that worked locally crashed after 200 pages in production. Memory leaked, sessions corrupted, and the whole thing needed to be rebuilt. Here is what we built instead.
Want to see how this was built?
View all projects
