Blog
Engineering write-ups from building AI products in production.
Why we chose DuckDB over pgvector for embedded semantic search at Smart Hire AI
An architecture decision write-up: we evaluated pgvector, FAISS, and DuckDB for our candidate-search vector index at Smart Hire AI. DuckDB won. Here's how we got there, the tradeoffs we accepted, and what we'd change in v2.
May 18, 2026
Hybrid LLM routing in production: when Ollama beats Groq (and when it doesn't)
Notes from production: we route between Ollama (local) and Groq (cloud) using confidence-gated fallback at Smart Hire AI. Here's the routing logic, the metrics that actually matter, and the failure modes that took us by surprise.
May 10, 2026
Building a 96%-accurate Hebrew OCR pipeline: lexical post-processing as the secret weapon
Notes from Introva: how we got from 78% to 96% accuracy on Hebrew historical text OCR. The model wasn't the bottleneck. The post-processing layer was. A 23,000-word lexical database did most of the heavy lifting.
April 28, 2026
Voice-first recruiting: how we built AI-led screening interviews on Vapi
An architecture deep-dive on Smart Hire AI's voice interview system. Vapi for the call layer, MediaPipe + TF.js for proctoring, transcript-driven rubric scoring, and the hard problems that show up only in production.
April 15, 2026
Per-student RAG: building AI mentors that actually know the student
How we built per-student vector stores at Smart Grade AI so the AI mentor surfaces real weakness patterns instead of generic advice. Architecture, gotchas, and what we'd change.
April 2, 2026