MLOps for independent research
Find out how working on an independent research project led me to apply my MLOps skills to create a performant and cost-effective experiment infrastructure
Alexandru Burlacu
A blog about advanced machine learning topics, MLOps, software engineering, distributed systems, and more.
Latest posts
A fable about MLOps… and broken dreams
A fable about a company's journey through scaling their ML function, and some practical advice on how you should do it
How to Solve the Model Serving Component of the MLOps Stack
It's important to be able to deploy a machine learning model when trained. But how do we approach serving ML models correctly?
Interviewing for a Senior ML Engineer position
My experience interviewing for a few Senior ML and MLOps roles. You will learn what are the common steps, quirks, and tips how to nail an interview for senior ML engineer positions.
AutoML Solutions: What I Like and Don’t Like About AutoML as a Data Scientist
AutoML sounds like magic. But how effective is it? And when to better use a simpler approach?
Choosing programming languages for real-world projects
How to pick a tool, language, or framework when real money and the business is at stake. What to consider when faced with this kind of situation.
Becoming a Senior Engineer
Some advice how to grow to a senior engineering role. What skills are most valuable for a senior software engineering career, and how to aquire them.
Going beyond simple error analysis of ML systems
When deploying machine learning algorithms, the stakes are much higher than in any toy problem or competition. For this reason, we need a much more thorough evaluation of our models, to make sure it is indeed good.
K-Means tricks for fun and profit
K-Means is an interesting, simple, and pretty intuitive algorithm. It turns out it can do more than just clustering, for example classification.
Logging, Tracing, Monitoring, et al.
When it comes to production-ready systems we need a way to know what's going on in it, aiding us in debugging it, when the time comes.