AVAILABLE FOR ADVISORY · Q3 2026

The architect behind enterprise AI that actually ships. actually works. actually scales.

I'm Santosh Varma — AI Architect at Kore.ai. I design the voice gateways, RAG pipelines, and agent runtimes underneath the AI products Fortune 500s actually deploy.

§ 01Positioning

Most teams can wire up a chatbot. Few can keep one honest, fast, and safe at Fortune-500 scale.

↑ That gap is where I work.

§ 02What I architect
i.

Voice AI
Gateways

Real-time conversations over the phone — ASR, TTS, barge-in, intent on top of SIP/WebRTC telephony.

SIPWebRTCASR/TTS<800ms
ii.

RAG & Hybrid
Retrieval

Retrieval that actually retrieves. Semantic + lexical, re-rankers, grounded-response eval at every step.

Q Hybrid Rerank LLM Answer
MilvusWeaviateVespa
iii.

Agentic
Systems

Autonomous agents that get real work done — tool-use, policy auth, recoverable state, runtime observability.

Agent Plan Act Verify Recover
Tool-usePolicyReplay
iv.

AI Platform
Engineering

The unsexy infra that lets the interesting work survive: routing, isolation, cost/latency budgets, eval pipelines.

Product Agents RAG Model Routing Eval & Obs
Multi-tenantSLOEval-CI

"I architect the production AI underneath the chatbot you'll eventually call when your bank app freezes — and I make sure the answer it gives you is the right one."

— SV, on the job description

14+years shipping
5products in prod
$10Mprogramme scale
3continents

Trusted by teams shipping AI at

Kore.ai · Storable · Savor Brands · Skillsoft · OPSTECH · TGQuest
§ 03Selected work
N° 001
2026 — present
AI Architect

Kore.ai

Reference architectures for voice AI, RAG, and agentic workflows across the XO Platform — AgentAssist, SmartAssist, SearchAssist, Voice AI.

Voice GatewaysRAGAgents
N° 002
2024 — 2026
Staff Software Engineer

Storable Auctions

Modernised a legacy monolith into event-driven microservices. Redesigned the Sitelink (FMS) integration end-to-end.

Event-drivenMicroservicesFMS
N° 003
2021 — 2024
Senior Fullstack

Savor Brands

Multi-party white-label commerce platform & a video-based patient-doctor telehealth product. Honolulu.

eCommerceWebRTC$10M scope
N° 004
Side projects
Architect & Builder

OPSTECH · TGQuest

Autonomous micro-agents for workflow automation. Hybrid retrieval, task orchestration, context-driven execution.

AgentsRAGAutomation
§ 04Live · AI clone

Try the work.
Right here, right now.

A small agent grounded on my résumé, capabilities, and case studies — running entirely in your browser. Recruiters: interrogate it. Founders: ask it how I'd approach your problem.

  • Why are you the right architect for enterprise voice AI?
  • RAG over 10M proprietary docs — how?
  • 3 bullets — why hire you?
  • Most under-rated production lesson?
sv.agent · v1.0 haiku-4-5
Sv

Hi — ask me anything about my work or how I'd tackle a problem you're stuck on. I'll keep it concrete and short.

§ 05How I work — five principles
  1. 01

    Honesty over hype.

    If the model can't answer it well, the system says so. Eval is a first-class part of the architecture, not a metric you add later.

  2. 02

    Latency is a feature.

    A 200ms answer that's "good enough" beats a perfect answer at three seconds — especially over voice. I design to the budget first.

  3. 03

    The boring infra wins.

    Routing, isolation, observability, recovery. The interesting work only stays interesting because the boring parts are bulletproof.

  4. 04

    Build for replacement.

    Every model, every provider, every chunking strategy gets swapped within 12 months. I architect for the swap, not the commitment.

  5. 05

    Mentor the next layer.

    A platform that depends on one person knowing how it works is a platform with a single point of failure. I leave teams stronger.

§ 06Receipts

Architecting AI-enabled voice gateways on the Kore.ai platform — ASR, TTS, SIP/WebRTC telephony, real-time streaming, and LLM-driven intent resolution powering enterprise voice agents.

— Current role · Kore.ai · 2026

Implemented process improvements resulting in a 20% reduction in issue resolution time.

— Mahathi Software

Led evaluation and enhancement of software/hardware interfaces, resulting in a 30% improvement in system performance and reliability.