Machine Learning Advanced
TRACE: Train Agents on Their Own Failures With GRPO
Diagnose an agent's missing skills, then train targeted, routed LoRA adapters with GRPO.
10 min read·Kodetra Technologies
TodayGuides
AI Guides for Builders
How-to content for builders, indie hackers, and AI engineers. Less theory, more shipped code.
Machine Learning 
Machine Learning Intermediate
TabFM: Zero-Shot Table Predictions That Beat XGBoost
Run Google's TabFM on real tabular data. No tuning, no feature engineering, one forward pass.
10 min read·Kodetra Technologies
Today
Machine Learning Advanced
Verifiers v1: Train Agents Past the Context Window
Ship a taskset, swap any harness, and turn compacted rollouts into real RL training samples.
13 min read·Kodetra Technologies
Today
Tutorials Intermediate
Context Compaction: Run Claude Agents Past 200K Tokens
Use Anthropic's compact-2026-01-12 beta so long agentic loops survive past the 200K context window.
11 min read·Kodetra Technologies
8d ago
Tutorials Intermediate
Run LongCat-2.0 in Python: Meituan's Owl Alpha
Meituan's 1.6T open MoE topped OpenRouter as 'Owl Alpha.' Call it in Python with the OpenAI SDK.
8 min read·Kodetra Technologies
8d ago
Tutorials Intermediate
Gemini 3.5 Pro: Feed a 2M-Token Codebase in One Call
Load a whole repo into Gemini 3.5 Pro's 2M context, query it without RAG, and cache to cut cost.
8 min read·Kodetra Technologies
13d ago
Tutorials Intermediate
Code Mode for MCP: Let Claude Write Code to Call Tools
Cut MCP agent context up to 99% by exposing tools as a code API the model calls in code.
9 min read·Kodetra Technologies
19d ago
Tutorials Intermediate
Stateless MCP in Python: The 2026-07-28 Handle Pattern
Port your MCP server to the stateless 2026-07-28 spec using the explicit-handle pattern.
10 min read·Kodetra Technologies
19d ago
Tutorials Intermediate
Progressive Skill Loading: 40+ Agent Skills, No Bloat
Build a skill-manifest registry so an AI agent wields dozens of skills without context bloat.
9 min read·Kodetra Technologies
19d ago
Tutorials Intermediate
Fable 5 Prompt Caching: Slash 1M-Token Codebase Costs
Reuse a huge codebase prefix across every Fable 5 call and pay ~90% less.
8 min read·Kodetra Technologies
28d ago
Machine Learning Intermediate
DiffusionGemma in Python: Generate Text 4x Faster
Run Google's open diffusion LLM with Transformers and learn why it decodes text in parallel.
9 min read·Kodetra Technologies
29d ago
Tutorials Intermediate
DeepSeek V4 Pro: Cheap 1M-Token Context in Python
Use DeepSeek V4 Pro's auto KV cache to run huge-context jobs for cents.
8 min read·Kodetra Technologies
Jun 13Page 1 of 2