EngineersAI & Agents
Serving LLMs on-premise: GPU, quantization, and throughput
A technical guide to sizing GPU and memory, applying quantization, continuous batching, and KV cache when serving language models inside the company.
5 min read
AI & Agents
8 publications
EngineersAI & Agents
Hybrid retrieval and per-user permissions in the index: how not to leak data in RAG
Lexical plus vector search with result fusion, source-inherited ACL applied at query time, revocation, multi-tenancy, and automated leakage tests.
5 min read
EngineersAI & Agents
RAG engineering: chunking, embeddings, and reranking
Structure-aware chunking, embedding versioning, reindexing, reranking, and retrieval evaluation — the technical decisions that determine RAG quality.
5 min read
ExecutivesAI & Agents
RAG is not magic: the real cost of giving AI enterprise context
Answer quality is a data and permission problem, not a model problem. What changes when the index must respect who can see what.
3 min read
EngineersAI & Agents
Continuous LLM evaluation: regression datasets that survive model swaps
How to build a regression dataset from real cases, pick metrics that matter, and use an LLM-as-judge calibrated against human review to block bad deploys in CI.
5 min read
EngineersAI & Agents
Prompts as code: versioning, testing, and evaluation
Prompts and system instructions treated with the same discipline as code: repository, PR review, automated tests, canary rollout, and rollback.
4 min read
ExecutivesAI & Agents
Prompt engineering didn't die. It moved up a layer.
What actually replaces the per-turn prompt in agentic systems — and where the risk goes when no one reviews the wording anymore.
9 min read
EngineersAI & Agents
Reference architecture for enterprise AI: layers, contracts, and perimeter
Model gateway, orchestration, retrieval, memory, identity, observability, and audit trail: the technical design behind an enterprise AI that isn't locked to a single vendor.
4 min read
Put it to work
Bring this into your company with the e.works platform
Everything we cover here — AI governance, legacy system integration, audit trails and preserved corporate knowledge — is available on the e.works platform at eworks.cloud.
e.works infrastructure on AWS
A managed environment protected by e.works on AWS, with encryption, per-company isolation, backup and high availability.
On-premises, in your environment
The same platform running in your company's data center or private cloud, when data sovereignty requires that nothing leaves your perimeter.
In either model your data stays yours — with access control, audit logging, configurable retention and guaranteed availability.
Newsletter
Technical and strategic content, once a month
Analysis on automation, industrial data and technology adoption. No spam.