Deep dives into AI infrastructure, API design, and the technical decisions behind Shok-IS.
🧠
Engineering
How We Achieved Sub-50ms Inference Latency at the 99th Percentile
A technical deep dive into our inference stack — from model quantisation to connection pooling, and the three architectural decisions that cut our p99 latency by 68% over six months.
SR
Sophia Reid · CTO·12 min read·March 2025
🔒
Security
Achieving SOC 2 Type II in 8 Months as a Startup
What we learned, what surprised us, and the exact controls checklist we used to pass our Schellman audit without a single finding.
AL
Amara Laurent · Security·8 min read
⚙️
AI & ML
Explainable AI in Production: Why We Return Vector Weights
Black-box predictions erode trust. Here's why every Shok-IS inference returns SHAP-style attribution weights alongside the prediction — and how customers use them.
SR
Sophia Reid · CTO·6 min read
📊
Product
Introducing Bidirectional Webhooks: The Full Story
March 2025 · 5 min
🌍
Engineering
Multi-Region Architecture: Lessons from Our London → Singapore Expansion
Feb 2025 · 9 min
🔑
Security
Why We Chose Argon2id for API Key Hashing
Jan 2025 · 4 min
📡
Engineering
OpenAPI 3.1: Why We Migrated and What Broke
Dec 2024 · 7 min
🏢
Company
Why We Incorporated in the UK Instead of Delaware
Nov 2024 · 5 min
📈
AI & ML
Model Drift Detection: How We Catch Accuracy Degradation Before Customers Do
Oct 2024 · 6 min
Stay in the loop.
Engineering articles and product updates, every two weeks. No marketing fluff.