AI systems for regulated contact centers: voice agents, real-time compliance, and LLM evaluation pipelines.
Verifiable output. Deterministic grading. Traceable failure modes.
Currently: Taking pilot engagements: compliance audits, eval-harness builds, and LLM QA pipeline architecture for collections, direct sales, and BPO verticals.
-
dental-desk — Voice agent prototype for clinic appointment scheduling: FSM (Finite State Machine), real-time STT/TTS, turn-taking, tool calls, multilingual support. Built as a client engagement pitch. Live demo
-
auditguard-mcp — A compliance-aware MCP server. Seven-step pipeline: RBAC, PII detection, policy enforcement, audit logging. 15-case eval, 100% pass. Live demo Blog post
-
LLM Deploy Cost Calculator — Production-grade GPU sizing, cost comparison, and break-even analysis for LLM deployment. Architecture-aware VRAM (GQA, MLA, MoE), throughput model, replica multiplier, pricing tiers. 49 model variants, 38 API plans. Cost estimates include uncertainty bands (self-hosted −15%/+30%, API ±15%). Validated against inference-bench measured numbers. Live tool Blog post
-
Inference Bench — Reproducible vLLM vs SGLang vs llama.cpp benchmark on NVIDIA L4 + A100 (via Modal). 125 runs across FP16, AWQ, reasoning (Qwen3-8B), and MoE (Qwen3-30B-A3B). May 2026 sweep with vLLM v0.20.1 + SGLang :latest confirmed L4 numbers stable; on A100, vLLM leads at all concurrencies — 4,762 tok/s Marlin at c=64 vs SGLang's 2,231 (+113%). Dashboard
-
Scrutiny — FDCPA/Reg F call transcript audit in 60 seconds. 12-rule rubric with verbatim evidence quotes and statutory citations. Dual-path evaluator. Live demo · Blog post
-
RegTriage-OpenEnv — An OpenEnv RL environment where the reward signal is auditor approval. 12 tasks, severity-weighted F1, auto-fail caps. Live demo · Blog post
-
yojana-sathi - A voice-based personal knowledge assistant that helps citizens understand and use Indian government policies in their regional language. Speak in any major Indian language, get answers in the same language.
Previously shipped LLM features to enterprise production at a speech analytics company serving regulated contact centers: automated quality scoring, real-time compliance assistant, conversational analytics engine. Self-hosted on customer infrastructure. Zero data egress.
rituraj.info: notes on production ML, compliance systems, and agent architectures.
Pilot engagements for mid-market collections agencies and direct-sales companies: FDCPA/FTC compliance audits, automated QA pipeline builds, LLM evaluation harness setup.