Agentic Evaluation and Benchmarks Cheatsheet
A concise, practical cheatsheet for evaluating AI agents, chatbots, zero-shot LLM features and image generators: metrics, criteria, rubrics, scoring, variables and statistics (confidence...
A concise, practical cheatsheet for evaluating AI agents, chatbots, zero-shot LLM features and image generators: metrics, criteria, rubrics, scoring, variables and statistics (confidence...
A concise, practical glossary of the acronyms used today across software development, engineering, operations, DevOps, security, data, MLOps, AI and support. For each one: what it means, the...
A practical cheatsheet for reviewing pull requests written partly or entirely by AI coding assistants and agents: why review is now the bottleneck, what the evidence says about AI code quality and...
A practical cheatsheet for Python 3.15, released October 9, 2026: explicit lazy imports (PEP 810), the frozendict and sentinel built-ins, unpacking in comprehensions, UTF-8 as the default...
A practical cheatsheet on psychological safety for engineering leads: what it is and what it is not, the research behind it (Edmondson, Google's Project Aristotle, DORA and Westrum's culture...
A practical cheatsheet on making AI agents safe to retry and resume: why agents are distributed systems, idempotency keys and how to implement them, retries with exponential backoff and jitter,...
A practical cheatsheet on position sizing for investors and traders: why size matters as much as selection, fixed fractional (risk per trade) sizing, volatility targeting and inverse-volatility...
A practical cheatsheet for cutting LLM API costs without cutting quality: where the money actually goes (input vs. output vs. cached tokens), prompt caching and how to structure prompts for cache...
A practical, research-based cheatsheet on recovery from work: what burnout is and is not, the effort-recovery and job demands-resources models, the four recovery experiences (detachment,...
A practical cheatsheet on git worktrees for running several AI coding agents (or humans) on the same repository at once: the mental model, the commands, naming and directory layouts, per-worktree...
An evidence-led guide to learning technical material faster: what retrieval practice, spacing and interleaving are, the key experiments and meta-analyses with their actual numbers, why rereading...
A practical cheatsheet for choosing between prompting, retrieval and fine-tuning, and for doing parameter-efficient fine-tuning well when you need it: how LoRA and QLoRA work, the hyperparameters...
The future of trading belongs to those who know how to make AI work alongside human intelligence. Hedge funds, family offices and independent traders are entering a new era where traditional...
A practical cheatsheet on investing a windfall all at once or in installments: what Vanguard, Schwab and academic research found, why lump sum usually wins and what dollar-cost averaging actually...
A practical cheatsheet for getting reliable JSON out of LLMs: JSON mode vs. structured outputs vs. strict tool use, the JSON Schema subset providers accept, Pydantic patterns, how constrained...
A practical cheatsheet for trunk-based development with feature flags: the rules DORA measures, short-lived branches and merge queues, branch by abstraction, the four kinds of flags, percentage...
A practical cheatsheet for defending Python and JavaScript projects against compromised packages: the attack patterns behind the 2024-2025 incidents, dependency cooldowns in npm, pnpm, Yarn, pip...
A practical cheatsheet for engineering managers redesigning a technical interview loop now that candidates use AI assistants: AI-allowed vs. AI-prohibited stages, what published companies do,...
Model Context Protocol (MCP) is an open standard by Anthropic that gives AI assistants a clean, universal way to connect to tools, data sources, and services. This article explains the...
A practical comparison of the four main Python type checkers in March 2026 (mypy 1.19, Pyright 1.1.408, ty 0.0.21 beta, Pyrefly 0.55 beta): how they differ on inference and unannotated code, the...
An evidence-led cheatsheet on procrastination for engineers and knowledge workers: what it is and is not, why it happens (mood repair, task aversiveness, temporal motivation theory), what it...
A practical cheatsheet for AGENTS.md: what it is, where the files go and which one wins, what to put in and leave out, a realistic example, a monorepo layout, how Codex, Copilot, Cursor, Gemini...
This article provides a concise reference for evaluating large language models (LLMs) and retrieval-augmented generation (RAG) systems. It covers core metrics like accuracy, F1, BLEU, ROUGE, and...
This article provides a concise reference for evaluating multi-step AI agents and agentic systems. It covers core metrics for task completion, reasoning, and efficiency, and highlights recent...
Modern AI systems may look diverse on the surface, but under the hood they rely on a small set of recurring architectural and training ideas. This article distills foundational concepts - ranging...
Modern data and AI products are built on a small set of recurring Python tools for data processing, visualization, interfaces, and APIs. This article provides a concise conceptual overview of...
A practical cheatsheet for writing a personal Investment Policy Statement: the sections that matter, how to turn goals and risk tolerance into targets and ranges, rebalancing rules backed by...
A practical cheatsheet for PostgreSQL 18: asynchronous I/O and its settings, uuidv7() for time-ordered keys, virtual generated columns as the new default, OLD and NEW in RETURNING, temporal...
What the research actually says about interruptions and focus: Sophie Leroy's attention residue, Gloria Mark's field and lab studies on fragmented work, Cal Newport's Deep Work, and Paul Graham's...
A practical cheatsheet on Anthropic's Agent Skills: what a skill is, the SKILL.md format and its frontmatter rules, progressive disclosure, bundled scripts and resources, where skills run (Claude...
A practical cheatsheet on the valuation metrics behind the AI bubble debate: Shiller CAPE, forward P/E, the implied equity risk premium (Damodaran), the Buffett indicator and earnings yield vs....
A practical cheatsheet for observing LLM apps and agents in production: the OpenTelemetry GenAI semantic conventions (gen_ai.* attributes, span names and kinds, metrics, and their development...
A practical cheatsheet on spec-driven development with AI coding agents: the spec, plan, tasks, implement loop, what GitHub Spec Kit and AWS Kiro actually do, spec-first vs spec-anchored vs...
A practical cheatsheet for Python 3.14: officially supported free-threaded builds (PEP 779), template strings (PEP 750), deferred annotations (PEP 649/749), multiple interpreters in the standard...
A practical delegation cheatsheet for engineering leads: the seven levels of delegation (Management 3.0 delegation poker), Oncken and Wass's "monkey" and five degrees of initiative, context not...
A practical cheatsheet on prompt injection and agent security: direct vs. indirect injection, Simon Willison's lethal trifecta, OWASP Top 10 for LLM Applications 2025 (LLM01 and friends), and the...
A practical cheatsheet on concentration risk in cap-weighted index funds: how cap weighting works, how concentrated the S&P 500 is in mid-2025 and how that compares with history, how to measure it...
A practical cheatsheet for running open-weight LLMs on your own hardware: quantization formats (GGUF k-quants such as Q4_K_M, GPTQ, AWQ, MXFP4), VRAM math for weights and KV cache with a Python...
A practical cheatsheet of decision tools with research behind them: separating decision quality from outcomes (Annie Duke), Gary Klein's pre-mortem, base rates and the outside view (Kahneman and...
A practical side-by-side cheatsheet for the three terminal coding agents of mid-2025: Claude Code, Gemini CLI and OpenAI Codex CLI. Install and auth, instruction files (CLAUDE.md, GEMINI.md,...
A practical cheatsheet on context engineering for LLM apps and agents: what goes into the context window, the write/select/compress/isolate strategies, how long contexts fail (poisoning,...
What the research says about cognitive offloading to AI, from the Google effect to the 2025 studies on critical thinking and ChatGPT, and a practical protocol for engineers to get the speed of AI...
A practical cheatsheet on drawdowns and risk metrics, using the spring 2025 tariff sell-off as the running example: drawdown, max drawdown, underwater periods, the asymmetry of losses, volatility,...
A practical cheatsheet for the retrieval half of RAG: chunking strategies and sizes, metadata, choosing embeddings, hybrid search with BM25 and reciprocal rank fusion, cross-encoder reranking,...
A practical cheatsheet on vibe coding vs. AI-assisted engineering: what Karpathy actually meant, where vibe coding is fine and where it is not, and the guardrails that make AI-written code...
A practical cheatsheet for DORA software delivery metrics: the four keys and the 2024 changes (failed deployment recovery time, rework rate, throughput vs. stability), the 2024 performance...
A practical cheatsheet for the modern Python toolchain: uv for Python versions, projects, lockfiles, tools and scripts; Ruff for linting and formatting; a clean pyproject.toml with PEP 621...
A practical cheatsheet for engineering leads on one-on-ones and feedback: what a 1:1 is for, cadence, an agenda template, questions by goal, feedback models (SBI, Radical Candor, feedforward),...
Why New Year's resolutions fade by February and what the research says about building habits that last: context-cued automaticity, how long it really takes (Lally et al. 2010 and the 2024...
A practical cheatsheet of LLM system design patterns based on Anthropic's "Building effective agents": the augmented LLM, prompt chaining, routing, parallelization, orchestrator-workers,...
A practical cheatsheet on reasoning models such as OpenAI o1 and DeepSeek-R1: what test-time compute and RL on chains of thought mean, when to use them instead of a chat model, how prompting...
A practical cheatsheet for the new-year portfolio review after a strong 2024: calendar vs. threshold vs. hybrid rebalancing, the 5/25 bands, drift math, tax-aware rebalancing with cash flows and...
Thinking and describing how Spotify's Discover Weekly leverages machine learning and statistical models to generate personalized music recommendations.
2024 was the year AI won two Nobel Prizes, learned to reason step by step, and prompted serious books and essays about life in a world where machines can do most of what we do. If our work is no...
A very concise description of habits and principles to help you achieve wealth over time.
This article uses a fruit basket analogy to explain key concepts in machine learning evaluation metrics, including recall, precision, and F1 score, making it easier to understand how these metrics...
If you've ever wondered how your favorite apps seem to know what you like, or how self-driving cars navigate the streets, you're in the right place. In this article, we'll break down the basics of...
Gallup says only 23% of employees worldwide are engaged, Amazon is calling everyone back to the office five days a week, and three quarters of knowledge workers already use AI. A practical guide...
The 2024 World Happiness Report shows the young growing less happy in North America and Western Europe, and The Anxious Generation has started a fierce debate about why. A guide to what the...
The European Parliament has approved the AI Act and the UN General Assembly has adopted its first resolution on AI. A practical guide to what AI ethics means for people who build and ship AI...
A jury convicted Sam Bankman-Fried, OpenAI's board fired and rehired its CEO in five days, hundreds of AI leaders warned of extinction risk, and governments signed the Bletchley Declaration and...
More than 224,000 tech workers lost their jobs in the first eight months of 2023, Gallup finds 59% of employees quietly quitting, and new studies show generative AI raising productivity in some...
The US Surgeon General has declared loneliness an epidemic, the longest-running study of adult life says relationships are the strongest predictor of a good life, and Finland tops the World...
Concise list of algorithmic techniques worth remembering.
An overview of neural networks and deep learning - introductory quick read-up, including their use cases, and pros and cons for each technique.
Github public repos list and resources.
Useful free software - tools and utilities to have handy.
Summary of the ideal use cases for models implemented in the scikit-learn Python library for machine learning.
10 points on time management from Seneca.
Proven by social psychology, studies have shown that these guide our behaviour.
Sequence of steps to consider for creating an effective persuasive presentation.
What are fallacies, categories and examples.
Concise points to be persuasive.
Key points from the teachings of Lao Tzu.
Studies available today show these are the keys to have a happy life.
A small curated list of Python libraries and frameworks in use for Data Science, Quantitative Development and other technical use cases. Of course, you may need any of these in other domains as well.
List and summary description of many Python libraries and frameworks that are used in the Data Science community in particular.
Current studies suggest that these are the most useful, impactful exercises you can do at home.
Based on the latest studies and research, these are the the keys.
An introduction to Machine Learning Classification, models, applicability and scikit-learn specifics to get an initial overview of what's possible.
Based on the latest studies and research, these is what a healthy diet means.
How colors affect purchases.
PyCon presentation on Python using Kafka.
Must have open source software anyone will have to use at some point.
Project Management quotes; to reflect, to analyze.
Starting with Docker.
Starting with the Python programming language.
Useful, concise tips, pointers and advice for your next Tech interview. A few more specific for Software Engineering interviews in particular also.
ChatGPT reached an estimated 100 million users in two months, Bing's chatbot had to be reined in within ten days, a science fiction magazine closed submissions under a flood of machine-written...
Collins Dictionary named "permacrisis" its word of 2022: a year of war in Europe, the highest inflation in decades, an energy crisis and a pandemic that has not quite ended. Stress and worry are...
A TikTok video in July turned "quiet quitting" into the workplace phrase of the year. Gallup says at least half of the US workforce fits the description, and Microsoft finds leaders and employees...
A Google engineer says the company's LaMDA chatbot is sentient; Google and most researchers say it is not. The debate about machine consciousness is fascinating philosophy, but it distracts from...
Americans quit 47.8 million jobs in 2021, a record. New research from Pew and MIT Sloan Management Review suggests that pay matters, but disrespect, lack of growth and toxic culture matter more...
Starting with docker-compose by giving an introduction and simple, concise examples.
Starting with Selenium and Python.
This autumn, leaked internal research showed that Facebook knew its ranking systems could amplify anger and that Instagram made body image worse for some teenage girls. In the same months, Oliver...
In 2021, large language models went from research curiosity to product. GPT-3 is available through an API, GitHub Copilot writes code, and Stanford has named the whole category "foundation...
Many people are not depressed or burnt out, just joyless and aimless. Sociologist Corey Keyes calls it languishing, and Adam Grant's April column made it the word of the spring. What the World...
A year after offices emptied, Microsoft's new Work Trend Index finds that 54% of workers feel overworked, 41% are thinking about leaving their employer, and company networks are becoming more...
Yesterday, 90-year-old Margaret Keenan became the first person to receive a clinically authorised COVID-19 vaccine outside a trial. For the next few months there will not be enough doses for...
In August, an algorithm lowered about 39 percent of teacher-assessed A-level grades in England, students marched on Parliament Square, and within four days the government abandoned it. The model...
Tomorrow's Clap for Our Carers is billed as the last. Ten weeks of applause have shown how much we value work that keeps others alive, and how rarely we say so. What research on job design,...
Quick, concise introduction on implementation of design patterns in Python, with example code.
The knowledge and understanding expected from a Software Architect.
Empty supermarket shelves make it look as if people turn selfish when things go wrong. Decades of disaster research say the opposite: most people cooperate, and the 405,000 volunteers who signed...
List of Python libraries we should use more.
Another beginning for useful content.
Videos of real-time reactions to what happened on May 6th 2010 - famous stock market flash crash.
The story of the new beginning.
Enabling swap on Linux.
How to programmatically retrieve the list of repos for a given user using the Github's API.
Time perspective.
Another beginning for useful content.
Rules of action.