Tutorials Intermediate
Kimi K3 Vision: Extract Structured Data From PDFs in Python
Use Kimi K3's native vision and JSON schema output to turn messy PDF pages into clean, typed data.
9 min read·Kodetra Technologies
YesterdayGuides
AI Guides for Builders
How-to content for builders, indie hackers, and AI engineers. Less theory, more shipped code.
Tutorials 
Tutorials Intermediate
Port Kimi K2 Agents to K3: The Reasoning-History Trap
kimi-k3 is not a drop-in swap. Map the params right and dodge the trap that breaks tool loops.
9 min read·Kodetra Technologies
Yesterday
Tutorials Intermediate
Kimi K3: Query a Million-Token Repo for Pennies
Point the OpenAI SDK at Moonshot's 2.8T K3, load a whole repo, and cut cost with caching.
8 min read·Kodetra Technologies
Yesterday
Tutorials Intermediate
Muse Spark 1.1: Drop-In Agent, Self-Managed Context
Point the OpenAI SDK at Meta's agent model, add tools, let it self-manage a 1M-token context.
10 min read·Kodetra Technologies
2d ago
Tutorials Intermediate
GPT-5.6 Persisted Reasoning: Reuse Thinking Across Turns
Use reasoning.context to reuse GPT-5.6's chain of thought across turns and cut redundant tokens.
9 min read·Kodetra Technologies
2d ago
Machine Learning Intermediate
Run Bonsai 27B on Your Phone: 1-Bit Local AI
Run the first 27B-class model on a phone: MLX, llama.cpp, tool calls, and the memory math.
10 min read·Kodetra Technologies
3d ago
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
4d ago
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
4d ago
Tutorials Intermediate
Prompt Cache Retention: Cut OpenAI Bills Up to 90%
Structure prompts, set prompt_cache_retention, and read cached_tokens to slash GPT-5.6 input costs.
8 min read·Kodetra Technologies
8d ago
Machine Learning Intermediate
SWE-Together: Grade Coding Agents on Multi-Turn Sessions
Build a SWE-Together-style multi-turn coding-agent eval with an LLM user simulator in Python.
10 min read·Kodetra Technologies
11d ago
Machine Learning Advanced
Read an LLM's Mind: Probe Hidden Thoughts Like J-Space
Train a linear probe on hidden activations and steer output, the method behind Anthropic's J-Lens.
9 min read·Kodetra Technologies
11d ago
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
12d agoPage 1 of 6