Engineering Manager / Technical Lead
A topic hub collecting every article tagged AI Engineering. Use it to explore related posts and follow this theme across the site.
19 articles
Frontier labs increasingly claim their models understand intent with less prompting. Here is what the evidence supports, what training changed, and how to test ask-versus-guess behavior in your own agents.
Compare Fable 5 Low, Opus 4.8 Low, and Sonnet 5 Low for coding and planning, including task cost, cross-effort surprises, and a practical routing policy.
Build a small Claude model router with explicit task classes, Low effort, bounded tools and budgets, isolated workspaces, acceptance tests, and cost-per-result evaluation.
Compare Fable 5, Opus 4.8, and Sonnet 5 at High effort using CursorBench 3.2, then see when Ultracode changes the model-routing decision.
How Google Translate evolved from rules and statistical phrase models to neural MT and today’s Translation LLM — and how to build a translation app in 2026 with the right mix of NMT, multilingual models, and LLMs.
For engineers evaluating gpt-oss-120b: OpenAI's open-weight model is fast for an open reasoning checkpoint, but throughput depends heavily on the provider and benchmark view. Here is how to read Artificial Analysis, OpenRouter trends, and the cost tradeoffs behind fast inference.
A source-backed review of Nous Research's Hermes Agent — what it is, how its memory and retrieval layers work, why it is not a classic RAG app, and which architecture decisions are worth borrowing.
A practical guide to Hermes scheduled agent jobs — daily digests, silent monitors, delivery targets, token governance, and why proactive WhatsApp cron hits Meta's 24-hour window.
How to deploy Hermes as an internal support copilot for Chatwoot agents — draft-only assist, private notes, tool scoping, memory governance, and quality gates beyond a thin ChatGPT wrapper.
A practical guide to wiring Meta's WhatsApp Cloud API through Chatwoot into Hermes Agent with OpenAI — architecture, bridge service, handoff to humans, and common operational pitfalls.
How to design a production Graph RAG app with graph construction, hybrid retrieval, durable hand-offs, evaluation gates, and a practical recommended stack for engineers and PMs.
A technical follow-up for engineers choosing embeddings, vector stores, rerankers, caches, and eval libraries for production MCP + RAG over public app data — with benchmark-backed comparisons and three reference stacks.
A practical feasibility guide for engineers and PMs on processing hundreds of subagent-style jobs, where the real limits are, and what RAG quality, evals, queues, and budget controls must exist before this is production-safe.
A grounded, motivating guide for students, juniors, switchers, and PMs on why CS fundamentals still matter, how AI skills layer on top, and what to build for the 2026 job market.
A friendlier roadmap for curious CS students who want to build real AI products now: what to learn, what to build, which job signals matter, and how to stay grounded while the market moves fast.
A practical guide for engineers and PMs designing RAG apps over well-structured, chained knowledge: how to chunk, retrieve, rerank, cite, evaluate, and improve with evidence instead of vibes.
A friendly, evidence-based comparison of Claude Sonnet 4.6 1M and Cursor Composer 2.5 Fast across speed, intelligence, coding, agents, cost, and product fit.
A practical engineering comparison of Claude Opus 4.7 1M alternatives: GPT-5.5, DeepSeek V4 Pro, DeepSeek V4 Flash, Gemini 3.1 Pro, Gemini 3.5 Flash, and Kimi K2.6.
A practical comparison of 1M-token and 200K-token LLM context windows: what gets easier, what still breaks, and how Engineers and PMs should choose an architecture.
Quick find
Search by topic, title, framework, or pattern.