Skip to content
HARBAN AI LABS
HARBAN AI LABS
Work with us

Search

Search

Find paper breakdowns, models, benchmarks, AI tools, services and guides across the site.

50 results for “evaluation”

Showing the 50 closest matches. Add a word to narrow the search.

Paper breakdowns

1 paper breakdown
  1. Attention Is All You Need: the architecture behind modern language modelsA technical reading of the 2017 paper that replaced recurrence with self-attention — what the Transformer does, why it trains faster, and what the paper did and did not show./research/attention-is-all-you-need

Benchmarks

3 benchmarks
  1. LiveCodeBenchCode generation (plus self-repair, execution and test prediction) on contest problems from LeetCode, AtCoder and Codeforces, tagged by release date./benchmarks/livecodebench
  2. MMLU-ProReasoning-heavy knowledge questions across 14 disciplines with ten answer options per question (~12K items)./benchmarks/mmlu-pro
  3. American Invitational Mathematics Examination (AIME), as used in LLM evalsSolving the 15 integer-answer (0–999) problems of a given year's AIME exam(s), typically averaged over repeated samples./benchmarks/aime

Articles

3 articles
  1. A practical checklist for shipping a RAG systemWhat to have in place before a retrieval-augmented generation system meets real users: a retrieval evaluation set, structure-aware chunking, hybrid search, verifiable citations, permission-aware retrieval, a freshness plan and production monitoring./insights/rag-system-shipping-checklist
  2. Building a portfolio that proves AI engineering skillHow to choose, evaluate, write up and package projects so that a reviewer can see you build systems that work and measure whether they do — plus how to scope freelance AI work and estimate it honestly./insights/ai-engineering-portfolio
  3. How to read an AI benchmark table without fooling yourselfEight questions to ask before you trust a row on a leaderboard: who ran it, which version, under what settings, whether the benchmark still separates models, whether the model has seen it, what the grader checks, how big the noise is, and which harness produced it./insights/how-to-read-an-ai-benchmark-table

Course sections

32 course sections
  1. 28.11 CI for ML projects: matrices, caches, GPU runners and LLM evaluationsGit & GitHub Mastery · Chapter 28: AI/ML workflows/learn/git-github/28-ai-ml-workflows#2811-ci-for-ml-projects-matrices-caches-gpu-runners-and-llm-evaluations
  2. 28.21 Interview questionsGit & GitHub Mastery · Chapter 28: AI/ML workflows/learn/git-github/28-ai-ml-workflows#2821-interview-questions
  3. 6.8 Empty commits: --allow-emptyGit & GitHub Mastery · Chapter 6: Commits/learn/git-github/6-commits#68-empty-commits---allow-empty
  4. 4. The repositoryGit & GitHub Mastery · Chapter 32: Capstone/learn/git-github/32-capstone#4-the-repository
  5. 29.2 The methodGit & GitHub Mastery · Chapter 29: Production Troubleshooting/learn/git-github/29-production-troubleshooting#292-the-method
  6. 28.16 What can go wrongGit & GitHub Mastery · Chapter 28: AI/ML workflows/learn/git-github/28-ai-ml-workflows#2816-what-can-go-wrong
  7. 14B.17 What a signature covers, and what verification provesGit & GitHub Mastery · Chapter 14B: Configuration, Aliases, Tags and Signing/learn/git-github/14b-configuration-aliases-tags-and-signing#14b17-what-a-signature-covers-and-what-verification-proves
  8. 1.2 The problem version control solvesGit & GitHub Mastery · Chapter 1: Fundamentals/learn/git-github/1-fundamentals#12-the-problem-version-control-solves
  9. 18.16 Seeing and managing rules: /rules, gh ruleset, gh apiGit & GitHub Mastery · Chapter 18: Branch Protection and Rulesets/learn/git-github/18-branch-protection-and-rulesets#1816-seeing-and-managing-rules-rules-gh-ruleset-gh-api
  10. 14A.4 Renames and copies are computed: -M and -CGit & GitHub Mastery · Chapter 14A: History investigation/learn/git-github/14a-history-investigation#14a4-renames-and-copies-are-computed--m-and--c
  11. 10. What you hand in, and how it is judgedGit & GitHub Mastery · Chapter 32: Capstone/learn/git-github/32-capstone#10-what-you-hand-in-and-how-it-is-judged
  12. 21A.4 Fork pull requests and the approval gateGit & GitHub Mastery · Chapter 21A: GitHub Actions security/learn/git-github/21a-github-actions-security#21a4-fork-pull-requests-and-the-approval-gate
  13. 28.14 Secrets in AI repositoriesGit & GitHub Mastery · Chapter 28: AI/ML workflows/learn/git-github/28-ai-ml-workflows#2814-secrets-in-ai-repositories
  14. 28.2 What belongs in Git, and what does notGit & GitHub Mastery · Chapter 28: AI/ML workflows/learn/git-github/28-ai-ml-workflows#282-what-belongs-in-git-and-what-does-not
  15. 1.11 The ten-command diagnosisGit & GitHub Mastery · Chapter 1: Fundamentals/learn/git-github/1-fundamentals#111-the-ten-command-diagnosis
  16. 20B.9 Self-hosted runnersGit & GitHub Mastery · Chapter 20B: GitHub Actions: delivery, runners, cost and debugging/learn/git-github/20b-github-actions-delivery-runners-cost-and-debugging#20b9-self-hosted-runners
  17. 14A.1 Why this mattersGit & GitHub Mastery · Chapter 14A: History investigation/learn/git-github/14a-history-investigation#14a1-why-this-matters
  18. 21A.1 Why this mattersGit & GitHub Mastery · Chapter 21A: GitHub Actions security/learn/git-github/21a-github-actions-security#21a1-why-this-matters
  19. 2.1 Why this mattersGit & GitHub Mastery · Chapter 2: The Mental Model/learn/git-github/2-the-mental-model#21-why-this-matters
  20. 4.1 Why this mattersGit & GitHub Mastery · Chapter 4: The Working Tree/learn/git-github/4-the-working-tree#41-why-this-matters
  21. 28.12 A professional AI project repositoryGit & GitHub Mastery · Chapter 28: AI/ML workflows/learn/git-github/28-ai-ml-workflows#2812-a-professional-ai-project-repository
  22. 20B.21 Interview questionsGit & GitHub Mastery · Chapter 20B: GitHub Actions: delivery, runners, cost and debugging/learn/git-github/20b-github-actions-delivery-runners-cost-and-debugging#20b21-interview-questions
  23. 17.17 Review practiceGit & GitHub Mastery · Chapter 17: Pull Requests/learn/git-github/17-pull-requests#1717-review-practice
  24. 14A.14 Set questions: --left-right, --cherry-pick, --ancestry-path, --reflogGit & GitHub Mastery · Chapter 14A: History investigation/learn/git-github/14a-history-investigation#14a14-set-questions---left-right---cherry-pick---ancestry-path---reflog
  25. 1.9 A guided first repositoryGit & GitHub Mastery · Chapter 1: Fundamentals/learn/git-github/1-fundamentals#19-a-guided-first-repository
  26. 27.19 What can go wrongGit & GitHub Mastery · Chapter 27: Open source and team workflows/learn/git-github/27-open-source-and-team-workflows#2719-what-can-go-wrong
  27. 27.17 Code review as a practiceGit & GitHub Mastery · Chapter 27: Open source and team workflows/learn/git-github/27-open-source-and-team-workflows#2717-code-review-as-a-practice
  28. 8.1 Why this mattersGit & GitHub Mastery · Chapter 8: Merge/learn/git-github/8-merge#81-why-this-matters
  29. 14A.30 Interview questionsGit & GitHub Mastery · Chapter 14A: History investigation/learn/git-github/14a-history-investigation#14a30-interview-questions
  30. 28.1 Why this mattersGit & GitHub Mastery · Chapter 28: AI/ML workflows/learn/git-github/28-ai-ml-workflows#281-why-this-matters
  31. 28.19 Version notesGit & GitHub Mastery · Chapter 28: AI/ML workflows/learn/git-github/28-ai-ml-workflows#2819-version-notes
  32. 8. How to run the simulationGit & GitHub Mastery · Chapter 32: Capstone/learn/git-github/32-capstone#8-how-to-run-the-simulation

AI tools

1 tool
  1. LangChain (incl. LangGraph)Open-source frameworks (LangChain, LangGraph, Deep Agents) for building LLM-powered agents and applications, alongside the commercial LangSmith platform for tracing, evaluation and deployment./tools/langchain

Services

6 services
  1. LLM & model evaluationPlanned — not yet offered. Task-specific evaluation suites and benchmark analysis that answer a concrete decision — which model to ship, whether a release regressed, where a system fails./services/llm-evaluation
  2. Reliability & red-team evaluationPlanned — not yet offered. Structured adversarial testing of LLM applications and agents — prompt injection, unsafe tool use, data leakage and failure under stress./services/reliability-red-teaming
  3. AI agent engineeringPlanned — not yet offered. Design and build of tool-using agents with explicit control flow, scoped permissions, evaluation and observability from day one./services/agent-engineering
  4. LLM application developmentPlanned — not yet offered. Production LLM features and applications — from API integration and prompt design to structured outputs, testing and deployment./services/llm-applications
  5. Technical research & advisoryPlanned — not yet offered. Literature reviews, model and vendor assessments, experiment design and technical due diligence for AI initiatives./services/research-advisory
  6. RAG & knowledge systemsPlanned — not yet offered. Retrieval-augmented systems over your documents and data, measured on retrieval quality and answer faithfulness — not demos./services/rag-systems

Projects

1 project
  1. The HARBAN platform: a research and evaluation publication built around provenanceThis website, written up as an engineering project: content as code, schemas that reject unsourced facts, server-rendered pages, local-first bookmarks and a deliberately small Postgres footprint for enquiries and the newsletter./projects/harban-platform

Pages

3 pages
  1. Benchmark methodologyHow every benchmark number on HARBAN is sourced, labelled and compared./benchmarks/methodology
  2. AI EngineeringOur engineering practice: reference architectures for agents, RAG and evaluation./engineering
  3. Work with usRegister interest, describe an evaluation need or propose a research collaboration./work-with-us

Model registry/models
Benchmarks/benchmarks
Paper breakdowns/research
AI tools directory/tools
Work with us/work-with-us