News
Why Your Betas Explode: The Hidden Geometry of Multicollinearity
3+ hour, 37+ min ago (1781+ words) Why your regression coefficients keep changing, and what geometry has to do with it. Last quarter, I was presenting a marketing mix model to a senior director. The slide showed two beta coefficients side by side: Linear TV at +2.4, Digital…...
How I Mastered Data Structures and Algorithms for ML (In 6 Weeks)
1+ day, 2+ hour ago (1272+ words) The strategies, questions, and process I used to ace coding interviews. Since 2024, I’ve passed over 90% of coding interviews, and it’s not because I am some sort of genius. The majority of coding interviews in the data science and machine learning…...
How I’m Making Sure My Analytics Career Doesn’t Get Eaten by AI
1+ day, 23+ hour ago (478+ words) The analytics career I signed up for five years ago doesn't exist anymore, and honestly, I am fine with that. When ChatGPT became a household conversation in 2022, I had a feeling that AI is overrated in the short-term and underrated…...
How Much Does It Actually Cost to Run a Local LLM? (Euros per Million Tokens, Measured)
2+ day, 2+ hour ago (1150+ words) I measured the actual GPU electricity for eight local models on one RTX 3090 — and the cheapest wasn't the smallest, nor the priciest the biggest. “It runs on my own GPU, so it’s basically free.” I’ve said that, you’ve probably said…...
I Built My Second ETL Pipeline. This Time, I Started Thinking Like a Data Engineer
5+ day, 22+ hour ago (1742+ words) Building a production-ready RSS pipeline with Python, Docker, PostgreSQL, and Kestra A few months ago, I decided I wanted to transition from data analyst to data engineer. Like many people starting out, I was overwhelmed by the sheer number of…...
RAG Was Always a Temporary Workaround. What is Next?
6+ day, 2+ hour ago (944+ words) Why the future of AI memory relies on persistent neural state, not vector databases. Look closely at your favourite RAG system. Very closely. We labelled this “memory.” A more honest name would be: a very elaborate, high-latency game of telephone…...
Behind the Scenes of Distributed Training and Why Your GPU Wiring Matters as Much as Your Strategy
6+ day, 23+ hour ago (1856+ words) A measured look at distributed training, from DDP and FSDP to the ZeRO stages in between, and why the wiring between your GPUs matters as much as the strategy you choose Training a model on one GPU is straightforward. You…...
Measuring Structure Stability of Econometric Models
1+ week, 1+ day ago (566+ words) The simplest most important idea for time series forecasting In this article I want to take a page out of a data scientist’s book and explore an important guardrail that the data science community has implemented as part of their…...
Stop Ranking Agent Configs by Average Score
1+ week, 3+ day ago (1758+ words) Best-worst comparisons, MaxDiff-style judging, and Plackett-Luce utility scores give agent teams a cleaner way to decide which configs to ship, prune, and route toward next. Every agent developer eventually hits the same question: which version should I actually ship? You…...
Assemble Each RAG Generation Prompt from a Base Prompt Plus the Rules Each Question Needs
1+ week, 4+ day ago (1284+ words) Enterprise Document Intelligence [Vol.1 #8B] – A fixed BASE, the rules each question needs, one registry: the dispatcher that turns a parsed question into a typed LLM call Generation is the fourth brick. A reader landing here can pick up the first…...