AINexus
Log inJoin

Long Reads

In-depth articles written by AI builders — tutorials, architecture and lessons learned.

  1. 01
    article·1d ago·2 comments

    Tutorial: checkpointed agent loops you can safely retry

    #Tutorial

    by @priya

  2. 02
    article·4d ago·2 comments

    Tutorial: schema-validated extraction with a retry loop

    #Tutorial

    by @leila

  3. 03
    article·4d ago·2 comments

    Lessons learned: the five retrieval bugs that reached production

    #Lessons Learned

    by @tomas

  4. 04
    article·4d ago·2 comments

    Lessons learned running GPU capacity for 14 internal teams

    #Inference#Lessons Learned

    by @diego

  5. 05
    article·5d ago·2 comments

    Tutorial: a diff-scoped AI reviewer that doesn't spam your PR

    #AI coding#Tutorial

    by @samuel

  6. 06
    article·5d ago·2 comments

    Lessons learned: 3,000 AI review comments, ranked by usefulness

    #AI coding#Lessons Learned

    by @samuel

  7. 07
    article·5d ago·2 comments

    Architecture notes: autoscaling GPU inference without melting the bill

    #Architecture#Inference

    by @aiko

  8. 08
    article·6d ago·2 comments

    Architecture notes: where to put the model boundary in your service

    #Architecture

    by @leila

  9. 09
    article·6d ago·2 comments

    Tutorial: hybrid search with pgvector and reciprocal rank fusion

    #Tutorial

    by @tomas

  10. 10
    article·6d ago·2 comments

    Architecture notes: one gateway in front of five model providers

    #Architecture#Inference

    by @nadia

  11. 11
    article·8d ago·2 comments

    Lessons learned from 11,000 agent runs that went sideways

    #Lessons Learned

    by @priya

  12. 12
    article·8d ago·2 comments

    Lessons learned after six months of routing 40M tokens a day

    #Lessons Learned#Inference

    by @nadia

  13. 13
    article·8d ago·2 comments

    Tutorial: build a streaming LLM gateway with per-tenant rate limits

    #Tutorial#Inference

    by @nadia

  14. 14
    article·11d ago·2 comments

    Lessons learned: what 18 months of eval data actually predicts

    #AI coding#Lessons Learned

    by @elena

  15. 15
    article·11d ago·2 comments

    Tutorial: serve an 8B model at 3x throughput with continuous batching

    #Tutorial#Inference

    by @aiko

  16. 16
    article·12d ago·2 comments

    Lessons learned from auditing 40 MCP servers

    #AI coding#MCP#Lessons Learned

    by @jonas

  17. 17
    article·13d ago·2 comments

    Lessons learned: 40k requests a day on a single 24GB card

    #Lessons Learned#Inference

    by @aiko

  18. 18
    article·16d ago·2 comments

    Architecture notes: evals, tracing and logs are one system

    #AI coding#Architecture

    by @elena

  19. 19
    article·16d ago·2 comments

    Tutorial: ship an MCP server your model can actually use

    #AI coding#MCP#Tutorial

    by @jonas

  20. 20
    article·16d ago·2 comments

    Tutorial: quantise a model without wrecking its reasoning

    #Tutorial#Inference

    by @marcus

  21. 21
    article·16d ago·2 comments

    Lessons learned maintaining an inference engine with 900 open issues

    #Lessons Learned#Inference

    by @marcus

  22. 22
    article·16d ago·2 comments

    Architecture notes: tool schemas as a product surface

    #AI coding#MCP#Architecture

    by @jonas

  23. 23
    article·18d ago·2 comments

    Architecture notes: a local-first inference stack for a team of five

    #Inference#Architecture

    by @marcus

  24. 24
    article·20d ago·2 comments

    Tutorial: scale GPU inference workers on queue depth with KEDA

    #Tutorial#Inference

    by @diego

  25. 25
    article·23d ago·2 comments

    Tutorial: a golden-set eval suite that blocks bad prompt merges

    #AI coding#Tutorial

    by @elena

  26. 26
    article·23d ago·2 comments

    Architecture notes: a citation-first RAG stack for regulated data

    #Architecture

    by @tomas

  27. 27
    article·25d ago·2 comments

    Architecture notes: one GPU node pool, three very different workloads

    #Inference#Architecture

    by @diego

  28. 28
    article·25d ago·2 comments

    Lessons learned: 6% of our first LLM outputs were quietly malformed

    #Lessons Learned

    by @leila

  29. 29
    article·27d ago·2 comments

    Architecture notes: planner–executor–verifier without the framework tax

    #Architecture

    by @priya

  30. 30
    article·27d ago·2 comments

    Architecture notes: what a code review agent should never read

    #AI coding#Architecture

    by @samuel