Alex Kim

I’m Alex, an AI engineer based in London, working on production AI/ML systems.

Most of my work sits where models stop being notebooks and start becoming product, infrastructure, automation, and decision-making systems. I care about the unglamorous part of AI: making systems reliable, measurable, and useful for real people.

What I Work On Now

At Intuit, my current work is around:

• automation systems that use AI to reduce manual work and support decisions
• AI features in customer-facing products
• AI agents, evaluation, and the systems needed to make them trustworthy

The interesting part is rarely just the model. It is deciding what should be automated, what still needs human judgment, how to measure quality, and how to make the result useful enough for real business decisions.

How I Got Here

• London, Intuit: working on production AI systems, automation, AI product features, agents, evaluation, and experimentation.
• Tel Aviv, Intuit: worked on ML-driven data aggregation systems, LLM evaluation approaches, and founded a 70+ member Data Science Guild.
• X5 Retail Group: moved into team leadership, working on large-scale retail forecasting, promotion optimization, pricing, hiring, and data quality.
• Kaspersky and Raiffeisen Bank: worked on NLP, support automation, compliance scoring, and customer identification.
• Earlier path: started as a data analyst, then moved toward data science because I wanted to build systems that could learn from data and affect real decisions.

Over 9 years in applied AI and ML, I’ve worked across LLMs and agents, NLP, automation, uplift modeling, forecasting, experimentation, and ML engineering. I’m also an inventor on 15+ patented systems and methods in applied AI and machine learning. The formal version is on my CV; this site is the more human version.

Why This Site Exists

This site is where I want to write down the practical lessons that usually do not fit into a CV.

Some posts will be technical. Some will be about career decisions, moving between roles and countries, becoming a lead and returning to IC work, and learning when the most useful work is not the most fashionable one.

I want to share particular stories, findings, and mistakes from my own experience: the kind of things that are hard to get from generic AI advice, but useful if you are building real systems or deciding what kind of AI/data career you want.