Blog

Field notes from the production frontier

Engineering deep-dives, opinion essays, and lessons from shipping AI for some of the world's most demanding companies. 25 articles and counting.

25Articles
19Categories
WeeklyNew posts

All articles

Browse the archive

RETRIEVAL · TOP-K
Retrieval

RAG that actually works in production

Retrieval is 80% of the answer quality and 5% of the demos. A field guide to chunking, reranking, citations, and the eight stages teams skip.

Mindlytic AI Team2026-04-02 · 16 min
600MS · TURN LATENCY
Voice AI

Why your voice agent feels off (and how to fix turn-taking)

Sub-second latency isn't the goal. Predictable latency, with the right barge-in semantics, is what makes a voice agent feel human.

Mindlytic AI Team2026-03-26 · 11 min
608271908894928698EVAL SCORES · WEEK 1–9
Evals

Evals, not vibes: how we ship LLM features safely

If you can't run a regression in CI, you don't have a feature — you have a demo. Building golden sets, judges, and drift detection.

Mindlytic AI Team2026-03-15 · 15 min
DEFENSE IN DEPTH
Safety

Defense in depth for enterprise AI

Prompt injection is the new SQL injection. Seven layers we put between the user and your most sensitive systems.

Mindlytic AI Team2026-03-08 · 13 min
Strategy

Buy, build, or assemble: a 2026 decision framework

The default answer to 'should we build it?' is no. The exceptions are where your moat lives.

Mindlytic AI Team2026-02-28 · 10 min
SENIOR ONLY · 5-PERSON SQUAD
Team

How to hire an AI team without building a research lab

You don't need ten PhDs. You need three senior engineers who have shipped one bad LLM product and learned why.

Mindlytic AI Team2026-02-20 · 9 min
ROI

Measuring AI ROI: a model that boards actually believe

Cost-per-resolution, not 'productivity uplift'. A four-quadrant model that links AI work to a P&L line.

Mindlytic AI Team2026-02-12 · 12 min
Models

Choosing a foundation model in 2026

Frontier vs. open-weights vs. fine-tuned small. A pragmatic decision tree, with cost numbers from real workloads.

Mindlytic AI Team2026-02-05 · 14 min
DURABLE WORKFLOW
Architecture

Agents vs workflows: when each one wins

If the steps are known, write the steps. Agents are for the unknown — and they're more expensive than you think.

Mindlytic AI Team2026-01-28 · 11 min
DEFENSE IN DEPTH
Compliance

The EU AI Act, in 12 things you actually have to do

We won't translate the regulation. We'll tell you what to put in your system, in what order, and which auditor will ask for it.

Mindlytic AI Team2026-01-20 · 13 min
Industry

AI patterns we see across fintech in 2026

Underwriting, KYC remediation, dispute triage, and one trap that every neobank falls into. Plus a benchmark of where the savings are.

Mindlytic AI Team2026-01-12 · 12 min
Industry

Where LLMs break in healthcare workflows

Charting, coding, and triage all look like text problems until you read the case law. A practitioner's view on the real risks.

Mindlytic AI Team2026-01-04 · 14 min
Cost

Designing for cost from day one

Token cost is the easy part. Idle compute, vector storage, and cold-start latency are where bills get out of hand.

Mindlytic AI Team2025-12-18 · 11 min
RETRIEVAL · TOP-K
Fundamentals

Embeddings, explained well — once

If you've read three blog posts and still aren't sure what a vector is doing, this one is for you. With pictures.

Mindlytic AI Team2025-12-08 · 9 min
600MS · TURN LATENCY
Voice AI

The voice AI stack in 2026

STT, VAD, dialog manager, TTS — and the four boxes you have to add if you want it to work in a contact center.

Mindlytic AI Team2025-11-29 · 12 min
Engineering

Prompt engineering, three years in: a postmortem

Some of it survived. Most of it didn't. What actually moves model behavior, and what was just superstition.

Mindlytic AI Team2025-11-19 · 10 min
Design

Designing the human in human-in-the-loop

Most HITL systems make humans into sad rubber stamps. Here's how to build review queues that humans actually want to use.

Mindlytic AI Team2025-11-10 · 11 min
Models

When to fine-tune — and how to know it worked

Fine-tuning is the expensive last resort, not the first lever. A checklist before you spend the GPU budget.

Mindlytic AI Team2025-10-30 · 12 min
Operations

Running an AI program office

How to triage 70 ideas down to 6 funded projects in 8 weeks, without crushing the people who proposed them.

Mindlytic AI Team2025-10-20 · 10 min
Data

Data readiness for AI: a 30-day audit

You don't need a lakehouse. You need three datasets cleaned, labeled, and exportable. Here's the audit we run.

Mindlytic AI Team2025-10-09 · 11 min
PLANNER
Case Study

Inside an agentic customer-experience deployment

An anonymized walkthrough of a deployment that took support contact rates from 18% to 7%, with the gotchas.

Mindlytic AI Team2025-09-28 · 13 min
608271908894928698EVAL SCORES · WEEK 1–9
Engineering

Observability for LLM systems

Your APM doesn't know what a hallucination is. What to log, what to sample, and how to find a regression at 3 a.m.

Mindlytic AI Team2025-09-17 · 11 min
Models

The quiet revolution of small models

Frontier models get the headlines; 7B fine-tunes do most of the work. Where small models are quietly winning in 2026.

Mindlytic AI Team2025-09-04 · 12 min
Procurement

The AI vendor RFP checklist

Twenty-three questions that separate vendors from poseurs. Print this, take it into your next meeting.

Mindlytic AI Team2025-08-22 · 8 min

Newsletter

One long-form piece in your inbox, every other Monday

No marketing, no recap of the news. Original engineering writing from the team that ships production AI for hospitals, banks, and insurers.

2,400+ subscribers · unsubscribe anytime

Want help shipping AI in production?

Talk to our team about a 2-week assessment.