SHEET 01 · PROFILE
Dr. Rajasekar Venkatesan
Architect of enterprise GenAI at scale. Translating frontier AI into production systems.
Work as a connected system
Profile ↗
01 / PlatformGenAI Service Layer
01 / PlatformAgent Service Layer
02 / ApplicationMulti-agent customer chatbot
02 / ApplicationAI travel-planning
03 / EssaySkills after compaction
03 / EssayWorkflows with Jev
03 / EssayMCP 2.0: What changed
- PracticeDr. Rajasekar Venkatesan
- 01 / PlatformGenAI Service Layer
- 01 / PlatformAgent Service Layer
- 02 / ApplicationMulti-agent customer chatbot
- 02 / ApplicationAI travel-planning
- 03 / EssaySkills after compaction
- 03 / EssayWorkflows with Jev
- 03 / EssayMCP 2.0: What changed
Enterprise AI architecture, research and practice.
Shared infrastructure serving 100+ production GenAI use cases.
A shared runtime for agent orchestration, memory, tools and governance.
Customer-facing agentic AI with 80%+ customer satisfaction.
AI travel-planning recognised with Gold at the Asian Design Awards 2024.
What Claude Code retains after compaction, how skills degrade, and patterns for restoring guidance while enforcing invariants outside the model.
A synthetic expense experiment separates bounded Jev judgments from exact calculations and policy rules, with explicit limits on what the results establish.
MCP field notes on durable state, principal-bound handles and retry correctness.
Explore the platforms, applications and ideas behind the work.
Explore the practice ↗Platforms → built-on applications · Practice → related essays
SHEET 02 · SIGNATURE WORK
Enterprise GenAI, in production.
What it looks like to run enterprise GenAI as a program, not a project.
GenAI Service Layer
When every team needs LLMs, you need infrastructure that turns chaos into reliability. Multi-cloud, multi-vendor, model-agnostic, with built-in routing, caching, observability, cost governance, and policy enforcement across 100+ production use cases.
Agent Service Layer
Agents aren't a feature you ship once. They're a class of system that has to be built, evaluated, validated, and operated. The framework standardises every step. Cuts the full pipeline (development, evals, human validation, load testing, deployment) from 16 weeks to 2 weeks or less.
Multi-agent customer chatbot
Customer-facing production AI handling 85K+ queries per week. CSAT lifted from ~50% to 80%+. A flagship example of what the Agent Service Layer can produce.
- Multi-agent orchestration with specialised sub-agents
- Parallel tool use and agentic delegation
- Human-in-the-loop escalation for high-stakes flows
- Persistent interaction + knowledge memory across turns
- Multi-step transactional workflows
Industry partnerships
Active engagement with industry-leading AI organizations: co-shaping enterprise GenAI roadmaps, evaluating frontier capabilities pre-release, and translating research into production.
Internal engagements
Building enterprise-wide AI fluency through training, upskilling, coaching, and mentoring across peers, engineering teams, business users, and senior leadership including C-suite stakeholders.
AI-powered travel-planning experience
Travel planning combines GenAI search with smart flight recommendations, built on the GenAI Service Layer. The experience connects shared AI infrastructure to measurable improvements in customer satisfaction and engagement.
- AI search: GenAI-powered search doubled customer satisfaction on the search experience.
- Flight recommendations: the recommendation surface averages 2,000 queries a day, with click-through rising from 23.8% to 34.4% in three months.
- Design recognition: the experience received Gold at the Asian Design Awards 2024.
SHEET 03 · WRITING
Latest essays.
What Claude Code Keeps of Your Skills After Compaction
A 5,000-token snapshot, a 25,000-token shared budget, and a skill index that never comes back. The mechanics, the failure modes, and the patterns that hold up in long sessions.
Building Complex AI Workflows with Jev
A live experiment in composing small judgments into larger decisions and knowing where the approach stops being useful
MCP 2.0? What Changed, What Remained, and How it Impacts Practitioners
Working notes on the 2026-07-28 revision of the Model Context Protocol, from someone responsible for running it in a large, regulated enterprise. Written for peers carrying the same responsibility.
SHEET 04 · HOW I THINK
How I think
Run GenAI as a program, not a project.
A one-off prompt is a feature.
Agents aren't a feature you ship once.
They're a class of system.
Prompts are technical debt.
Every prompt you write is a contract with a model version.
SHEET 05 · NOW
Currently working on.
The initiatives shaping the next chapter of enterprise GenAI at scale.
Agentic systems at scale
Pushing the Agent Service Layer into more departments. New eval harness, new tool primitives, faster idea-to-prod cycles.
Voice AI for internal ops
Native end-to-end streaming voice replacing legacy STT-text-TTS for internal staff workflows. Real-time, low-latency, prosody-aware.
Frontier model partnerships
Pre-release evaluations and roadmap co-shaping with OpenAI, Anthropic, Google, and AWS.
DocIQ + Tacit
DocIQ: hosted app deployed. Page-grounded document Q&A: lexical search, agent-directed exploration and visual inspection in a reusable retrieval engine. Tacit: local application. Scattered source material becomes useful organizational knowledge when people can trace, review and maintain it.
SHEET 06 · INPUTS
Reading & thinking.
A short list of what's shaping how I think this season.
Co-Intelligence
Ethan Mollick on living and working with AI. The clearest framing of human + agent collaboration I've read.
Anthropic's “Building Effective Agents”
The architecture patterns that hold up in production. Routine reading for anyone shipping agents.
Claude Code & Codex
The current state of AI coding agents. Reshaping how software gets written, not just how it's typed.
SHEET 07 · GET IN TOUCH
Open to collaboration, advisory, and speaking.
For partnerships, technical advisory, conference invites, or to compare notes on enterprise AI.