MatrixTradingSystem
Pushed the system from just runnable to actually testable for users. Fixed many real-world issues: broker integrations, order state synchronization, Windows installers, macOS releases, error diagnostics, log bundles, and various preflight checks. You really can’t see these things in a demo; they blow up while testing, and you fix them on the fly. I can only say the Windows environment is truly magical, and while some of these financial brokers are quite modern, others are very old-school. I hope the flag I set can be reached. For the next two weeks, my main focus will be getting this project done and shipping it!
Email operations command center
A tool to help enterprises handle email workflows. This was the topic our group chose in the AI PM course. At first, I felt this wasn’t a pain point of mine, so I didn’t spend too much thought on it. After the team leader sent the first version of the PRD, I figured I had to build it myself to really get a feel for it. So I started with Gmail integration, email syncing, AI reply drafts, case tracking, and reminder mechanisms, and later moved on to handling deployments, databases, Sentry, cron jobs, and rate limits. This line moved very fast, and it was obvious that there’s a big gap between a PoC and a true production launch. The process turned out to be quite interesting and not as boring as I imagined. Maybe in the future, we won’t even need to open Gmail or Outlook, and can operate entirely within our own tool.
Auris
An embedded voice AI device experiment. From ESP32 firmware, audio, and WebSockets to a Rust voice server, connecting to the OpenAI Realtime API, and finally adding a manager API, device registration, and token revocation. This is a very fun track because it’s not purely software—it’s the worlds of software, hardware, and AI glued together. However, I’ve been doing this project on and off, saying ‘Yes’ to a lot of steps along the way. I’ll have to circle back to this one after getting MTS shipped!
In the meantime, I also continued to expand and organize my personal knowledge base, covering AI platforms, embedded AI, solo dev workflows, and decision architectures. I also built some English test practice tools for my daughter, and quickly put together an AIACTL literacy test website for classmates to practice on.
To sum up, these two weeks weren’t just about writing a lot of code, but practicing one thing: how to use AI as a lever to quickly push an idea to something that can actually be tested, used, challenged, and then refined closer to a real product.
On a side note, I received this toy, what should I do now… a stepping stone into the robot world!
Comments & Feedback