Pulumi for AI Voice Infra: TypeScript + ESC + Pulumi AI (2026)
Provision an AI voice stack with Pulumi 3.230+ in TypeScript: ESC for secrets, AI-generated modules, deployments-as-a-service, and uv-backed Python providers.
11 min read
Agentic AI & LLMs
Agentic AI & LLMs archive page 354 of 631
Agentic AI & LLMs
Agentic AI, LLM engineering, and the models behind modern automation — multi-agent systems, LLM evaluation and comparisons, RAG, fine-tuning, AI infrastructure, security, and production AI engineering.
From the blog
Latest articles
Chat Agents for Toast, TouchBistro, and Square Restaurants: The 2026 Guest Service Layer
Toast IQ, TouchBistro AI, and Square's restaurant POS all ship AI in 2026 but lack a true conversational guest layer. Here is how a chat agent on top cuts phone-order errors by 60% and lifts off-premise revenue 18%.
10 min read
Hidden Costs in AI Voice: ASR, TTS, LLM, Telephony (2026)
The advertised $0.05/min is rarely your bill. We dissect the five real cost layers — ASR, TTS, LLM, platform, telephony — with current 2026 unit prices and an all-in worked example.
11 min read
Tool-Using Agent vs Orchestrator Split: The 2026 Decision Tree
Should the planning brain also call tools? Almost never. We separate orchestrator from tool-using workers, show why mixing the two is the #1 cause of unreadable traces, and how CallSphere enforces the split across 37 agents.
10 min read
Postgres RLS for Multi-Tenant AI: Tenant Isolation Done Right (2026)
Move tenant filtering out of your app and into Postgres. Row Level Security gives you database-enforced isolation that survives ORM bugs, hand-rolled SQL, and AI agents writing queries on the fly.
12 min read
Tuning VAD for AI Voice Agents: Thresholds, Hangover, Hysteresis (2026)
VAD doesn't make ASR more accurate — it controls endpointing latency, barge-in feel, and turn-taking. We tune Silero VAD threshold, prefix_padding, and silence_duration_ms with real production traces.
11 min read
SRTP Key Rotation in 2026: DTLS 1.3, Perfect Forward Secrecy, and Re-Keying
DTLS 1.3 (RFC 9147) is rolling out across browsers in 2026. SRTP keys must rotate per session and the cipher floor moved to AES-128-GCM with PFS. Here is what production HIPAA voice needs.
11 min read
RLHF for Chat Agents in 2026: PPO Is Out, GRPO Is In
The 2022 RLHF recipe (SFT → reward model → PPO) is largely dead in 2026 — frontier labs have moved to GRPO and DAPO with verifiable rewards. We unpack what changed, when reward modeling still matters, and how OpenRLHF makes it production-ready.
11 min read
Build a RAG AI Agent with Drizzle ORM + pgvector + OpenAI (2026)
Drizzle ORM ships first-class pgvector helpers in 2026. Wire HNSW cosine search to gpt-4o-mini for sub-50ms RAG with type-safe queries — no Python, no Pinecone.
11 min read