Skip to main content

Agent logs, traces-a padikkradhu

Agent-oda ovvoru step-aiyum log pannalaam, failed run-a padichu enga, yen thappaa pochu-nu kandupidikkalaam.

Intermediate 3 · Agent builderabout 3 hours7 sections

Short-aa sonna

Agent logs (traces) oru run-la irukkura ovvoru model call-aiyum tool call-aiyum inputs, outputs, errors, times, tokens, cost-oda record pannum. Failed run-a padichu, thappaana outcome-la irundhu pinnaadi nadandhu poi, enga thappaachu-nu kandupidikkalaam.

Idhu yen mukkiyam

Agent thappaa nadandhaa, logs illaama andha thappa fix panna mudiyaadhu. Tracing theriyum-nu solla mudinjaa, production-la agents-a handle panra engineers-a team virumbum.

Sign up free to read the whole lesson

The rest of this lesson: 7 parts, video lessons, the best course, practice and a quick check. It is free, takes a few seconds and needs no card.

Accounts are for learners aged 18 and over. Under 18? The quick answer and the first part of every lesson stay open: browse lessons by topic.

Lessons on Generative AI & LLM Engineer0 of 65 done
  1. 1What AI, machine learning and a language model are
  2. 2Use a chat assistant for a real task and check the answer
  3. 3Use AI safely at work
  4. 4Set up Python and run your first file
  5. 5Variables, loops and functions in Python
  6. 6Write your own functions
  7. 7Lists and dictionaries
  8. 8Files and errors
  9. 9The terminal and Git
  10. 10Packages and virtual environments
  11. 11Call a web API and read JSON
  12. 12Tables with pandas
  13. 13Call an LLM API from code
  14. 14System prompts and prompt design
  15. 15Tokens and the context window
  16. 16How a model picks its next word: temperature, top-p and top-k
  17. 17Get structured JSON output
  18. 18Errors, retries and cost
  19. 19How a chat model is made: pre-training, SFT and preference tuning
  20. 20AI in Indian languages: tokens, quality gaps and what to use
  21. 21Embeddings
  22. 22Chunk documents for retrieval
  23. 23Vector search
  24. 24Build a RAG app
  25. 25Check whether retrieval works
  26. 26Parsing documents and OCR
  27. 27Metadata that helps retrieval
  28. 28Hybrid search, reranking and filters
  29. 29Query rewriting, HyDE and parent-child chunks
  30. 30Grounding, citations and refusing to guess
  31. 31Long context or RAG
  32. 32Tool calling
  33. 33The agent loop
  34. 34Logs and traces for agents
  35. 35How agents fail
  36. 36Long documents without vectors
  37. 37Test sets and evals
  38. 38LLM judges and their limits
  39. 39Choosing a model: benchmarks, leaderboards and your own tests
  40. 40Scoring open-ended answers: exact checks, similarity and side-by-side comparison
  41. 41Catching hallucinations: why models make things up and how to detect it
  42. 42Regression tests for prompts
  43. 43Defend against prompt injection
  44. 44Reasoning models and test-time compute: when to let the model think longer
  45. 45Data pipelines for AI
  46. 46Permission-aware RAG
  47. 47Graph RAG
  48. 48Multimodal RAG
  49. 49How fine-tuning works: data, training and what can go wrong
  50. 50LoRA, QLoRA and the GPU memory maths
  51. 51Fine-tuning in practice: tools, settings and reading the loss curve
  52. 52Synthetic data and distillation: training with model-made data
  53. 53Building a training dataset: quality, coverage and quantity
  54. 54Deploy and monitor an LLM app
  55. 55Caching and speed
  56. 56Inside LLM inference: prefill, decode and the KV cache
  57. 57Grow an AI app in layers: context, guardrails, routing, caching and agents
  58. 58Advanced harness engineering
  59. 59Harnesses for long-running agents
  60. 60Cost control
  61. 61Security and privacy
  62. 62Prompting, RAG or fine-tuning
  63. 63Design an AI system
  64. 64Quality, cost and speed trade-offs
  65. 65The design interview

On these roadmaps: Generative AI & LLM Engineer, Forward Deployed Engineer (FDE), Agentic AI Engineer, MLOps Engineer

AgentsIntermediate 3: Agent builderVideo lessonsTracingAgentSpanLogging

Track

Today's AI turns are used

Lessons, roadmaps, videos, practice cards and the code playground stay free with your account, with no limit.