Instrumenting Coding Agents
Yapping about adding an observability layer to coding agents
I am Clelia, a passionate developer advocating for AI and search tech at Qdrant
I am a pythonist and a rustacean, coding geek and a strong open source advocate!

About Me
I am versatile, I am a builder: I love creating things, digging deep into code - here are some of things about me!
I am fully-proficient in python, the first language I fell in love with and the first one I wrote production code in. I use python mostly for backend services and AI agents.
I love Rust and I've been coding almost exclusively in it to get as proficient as possible. My project span CLIs, web servers, AI and DevTooling
Although I use it less than other languages, I am a great fan of TypeScript elegance and type system. I enjoy mostly building full-stack apps using TS for backend and frontend, as well as AI-related products.
I use bash daily for scripting and automation: I love creating ways to speed up repetitive workflows or to combine AI tools in self-supervisioning loops.
If I am not building things, I really enjoy writing about them: being it the latest advancements in tech and science, or a tutorial about one of my code projects, writing will always be part of my journey!
I am trying to get more and more into systems programming, learning lower level concepts and exploring programming languages such as Zig and C#. Will report back on how it goes!
Step 1: Designing
The first thing I do is to create a mental schema of what I would like to produce: the features, the tech stack, the intended audience... Everything needs to be drafted in the back of my mind.
Step 2: Learn and Research
Once I designed the idea, I start researching innovative ways to implement it, and I try to learn all the needed concepts. I mostly use good ol' browsers to do that, but sometimes I let LLMs help.
Step 3: Build
Once I have the design and the concepts, building starts. I try to code my applications hierarchically: for example, if I need to create a full stack app connected to an API, I first lay out the utilities, then the API/backend and finally the frontend.
Step 4: Iterate and fix bugs
After building, I test the behavior of my project and I try to find as many bugs as I can, fixing them after every failed iteration, until I have a satisfying result.
Ready!
Once I am satisfied, I prepare a nice README for my project, give it a license and contribution guidelines, and then I just push it out live - open source as always!

Beyond tech
Because when you're a dev, you don't just build tech, you need to live it
One of my core skills is being able to pitch projects in front of people by preparing visually appealing presentations to efficiently deliver my ideas.
Critical thinking is a skill that I cultivated in my life by taking part into competitive debates: I like to reflect, understand, question, until I get real answers.
I like things being transparent, and I feel like every good working relationship is built upon open and honest communication, a best practice that I always try to follow.
The blog contains articles related to scientific discoveries, as well as more technical articles about programming. Enjoy!
Yapping about adding an observability layer to coding agents
Commentary of my ongoing journey into building a Rust native addon for Node
A practical comparison of Rust and Go, with an overview of DevX and performance
Exploring the building blocks of LobsterX, a document-processing agent inspired by OpenClaw
If you wish to chat and know me professionally, make sure
to connect with me on LinkedIn!