Jakub Misiło

After joining Campy, I improved the test setup and code quality by introducing integration tests for search functionality and moving the platform into a monorepo.

More importantly, it allowed Campy to have centralized DB services, which allowed us to protect the data model (which was quite unstable, due to Firestore usage, which allows to write anything to the DB) with Zod contracts. This change improved the security and stability of the app significantly. Additionally, it helped to adapt to AI agents well, by sharing the context of the entire backend with the agents.

The next step was introducing the Content Machine: a content-processing workflow that pre-moderates, enriches (via a team of scrapers), and reviews (automatically, with AI and guidelines) new and updated data points such as locations, reviews, and images, with the possibility to extend it to updating existing data points by enriching them with new information. It helped to increase data accuracy, lift content quality, lower the load on the moderation team, and reduce manual ops. Additionally, it exposed several ghost errors that we were not aware of before, and led to a larger data model fix project, which was an extension of Zod contracts, mentioned before, that reduced the error rate to almost 0%.

Part of my work was also the development of the Analytics MCP server, which was designed to let non-technical team members connect it to their AI agent of choice and query the data warehouse directly, without any SQL knowledge, nor the need to reach out to the engineering team, due to a lack of skills & context about the data structures there. Despite it being a side project, developed over the off-site hackathon, it was a great opportunity to learn how to build MCP servers from scratch.

During my time as a founder of knmstudio, we worked on several landing pages & product projects. We helped early-stage founders of companies like Solg, Campy or Blooio with the design & development of their products & landing pages. It was an amazing experience to lead client projects from the very beginning, from acquiring clients to designing and developing the products.

We also spent our time on internal projects, such as Pagey and AI SDK Directory. The Pagey got its traction via PH launch & by being featured twice by the Superhuman AI newsletter!

I started my journey at New Native as a Data Science Intern. My first project was an internal Discord Bot, powered by GPT-3. It was supposed to enhance our day-to-day work. Then I quickly transitioned into a Software Development role and moved on to the lablab.ai platform product team, where I was responsible for building a Discord integration that served as a bridge between the lablab platform and the Discord. It connected the lablab account with Discord, automatically added users to server, created team channels and added team members to them, and more.

I was also responsible for the lablab AI dashboard bot, that was supposed to walk users through the entire event - from onboarding, finding & creating a team, to project idea generation and more.

Together with the team, we planned, designed & developed the MVP of the enterprise.lablab.ai platform for the Cohere x NYSE event. The platform served the event for Fortune 500 companies, and was successful!

We got an AI Startup of the Year 2023 award from The Europas Tech Startup Awards!

My last project there was a rebuild of core lablab.ai, which includes software architecture redesign, migration to monorepo, migration to a new CMS system, and a new API architecture.

Over the time I spent at NN, we had over 200k users on the platform that I've built from day one.