Agent platforms at Eightfold AI. Semantic search at Walmart.com. Conversational AI at Amazon Alexa. Machine translation at eBay that added $1B+ in GMV. I advise CEOs, CTOs, founders, and investors on what to build, how to build it, and what it is worth.
Choose your seat at the table. Each engagement draws on the same track record: shipped products, built teams, and raised capital.
The obstacle is rarely the model. It is strategy, platform, governance, and organization — and those are leadership problems.
Pilots multiply, but few reach production — and fewer show up in the P&L.
Quality, safety, and compliance are assessed once customers are already exposed, not before release.
Each team builds its own AI stack, multiplying cost and slowing every subsequent launch.
Delivery slips while competitors use AI to ship faster with the same headcount.
Engagements are fixed-fee and scoped to outcomes. Begin with a strategy engagement, or bring me in to lead execution.
The place most clients start. In two weeks I interview your leadership, review your architecture, data, and pipeline of AI work, and identify where AI will move revenue, margin, or cost to serve — and where it will not.
I lead your team to one shared platform for agents and search — orchestration, memory, tool access, retrieval, guardrails, tenant isolation, and evaluation gates — and ship the first use case on it.
For venture funds and acquirers: an operator’s assessment of an AI company — architecture, data advantage, evaluation rigor, inference economics, and team — before you commit capital.
For AI companies and their investors: ongoing counsel on technical strategy, senior AI hiring, and enterprise readiness — as an advisor or board member.
AI systems I have architected and led to production — agents, workforce intelligence, search, translation, and conversational AI — used by millions of people.
A shared runtime and authoring layer for agents — orchestration, memory, tool access, guardrails, tenant isolation, and a built-in evaluation layer — so several product lines ship agents on one foundation instead of one stack per team.
An agentic analyst for HR and business leaders: plain-language questions about skills, readiness, and talent become answers with the analysis behind them — no analyst queue.
An agent that carries a candidate from first search to offer letter over weeks of text conversation, with human checkpoints at every consequential decision.
A coaching assistant grounded in a skills knowledge graph, with long-term memory that carries each person’s goals and context across sessions.
Model routing, retrieval, fine-tuning, inference, and agent workflows on multi-tenant infrastructure for defense and education customers, certified to SOC 2 Type II.
Offline and online evaluation that gates every model and agent change before it reaches users — applied to voice assistants, search, and enterprise agents.
Query understanding — entity extraction, intent classification, and query rewriting — for one of the largest product catalogs in the world, with +85% long-tail understanding.
Generation, retrieval, relevance, and safety models for a consumer voice assistant, doubling user satisfaction while cutting cost to serve ~15%.
A production translation engine that opened cross-border marketplaces to buyers in their own language, adding $1B+ in GMV.
Language, search, and agentic systems — built and operated at enterprise scale.
Every engagement follows the same sequence, so investment decisions rest on evidence rather than enthusiasm.
Leadership interviews and a review of architecture, data, and delivery. We agree on the business number AI must move and measure today’s baseline.
Architecture, build-versus-buy decisions on models and vendors, and one scoped pilot with success criteria signed off by its executive owner.
The pilot ships to real users behind evaluation gates, and its impact is reported against the baseline every week.
The platform, playbooks, and hiring plan that let your team repeat the result across products without me.
The questions I bring to technical due diligence — from someone who has raised capital, built the platforms, and hired the teams.
What remains valuable when frontier models improve next quarter: proprietary data, workflow depth, distribution.
Whether the system can scale in users, tenants, and use cases without a rewrite — and where the debt sits.
How quality is measured and gated before release. Teams without evals are shipping demos.
Inference, retrieval, and human-review cost per outcome, and how it trends as usage grows.
Security, tenant isolation, compliance, and private deployment — what large customers will demand in procurement.
Whether the technical leadership can hire, ship, and make sound build-versus-buy decisions at the next stage.
Selected results from leadership roles. Further detail is available on request.
I keep a small roster so that every client receives senior attention.
No vendor referral fees or reseller arrangements. Recommendations serve your interests alone.
Under NDA from the first substantive conversation. Client work is never shared without permission.
You work with me directly, from the first assessment to production. No hand-offs to junior staff.
Success criteria agreed up front with an executive owner, and progress reported against a baseline.
I have spent 15+ years building AI that real people use — and the last decade leading the teams and companies that build it.
As Head of AI & Engineering at Eightfold AI, I lead four product lines and 50+ engineers. I architected the shared platform our agents run on and took three agentic products to production — agents that work for weeks alongside people, assistants with long-term memory, and analytics that answer leaders’ questions in plain language. As co-founder and CPTO of MarkovML, I raised $8M, built a 35-person company, and delivered an enterprise AI platform to SOC 2 Type II. Before that I led conversational AI at Amazon Alexa, founded the Query Understanding team behind Walmart.com search, and was the second engineer on eBay’s machine translation engine.
I work with CEOs and boards on where to invest, with CTOs on how to build, with founders on becoming enterprise-ready, and with investors on which AI companies will last.
Member of the extended leadership team, reporting to the CTO. Lead four product lines and 50+ engineers across the U.S. and India, with ownership of headcount, budget, and vendor spend. Architected the shared agent platform — runtime, authoring, memory, tool access, guardrails, and evaluation — and took three agentic products from first code to production, spanning long-horizon hiring agents, career guidance, and conversational workforce analytics. Made AI-assisted engineering standard practice across every team.
Led a 35-person company with engineering, platform, security, and go-to-market reporting in. Raised $8M, sold into defense and EdTech, achieved SOC 2 Type II, and presented strategy and risk to the board. Merged in 2025.
Doubled user satisfaction and reduced cost to serve by ~15%. Delivered the evaluation system that gated every model change before production.
Raised pre-seed funding and launched a multimodal conversational shopping assistant. Wound down when COVID closed retail, returning remaining capital to investors.
Founded the Query Understanding team behind Walmart.com semantic search, improving long-tail understanding by 85%. Led the replacement of Oracle Endeca across a 100+ engineer organization, reducing infrastructure cost ~13%.
Second engineer on eBay’s machine translation engine ($1B+ added GMV); Office for Mac storage; production software for AMD’s APU line.
With a 30-minute conversation about the outcome you need. If there is a fit, you receive a one-page proposal with objectives, measures of success, and scope within 48 hours. Most clients begin with the two-week AI Value Diagnostic.
Both. I brief executives and boards, and I review architecture and work directly with engineering teams to see recommendations through to production.
Every engagement is covered by an NDA. Client names and details are never published without written permission.
Yes. I accept no referral fees or commissions from model, cloud, or software vendors. I work across frontier, open-source, and private models on AWS, GCP, or Azure.
Yes. I run technical due diligence on AI companies for funds, advise portfolio companies on architecture, hiring, and enterprise readiness, and take a small number of advisory and board roles.
I take on a small number of advisory engagements that do not conflict with my role at Eightfold AI, and I decline work for companies that compete with it.
A small number at a time, so each engagement receives senior attention. Early contact is recommended if you have a deadline.
Whether you lead a company, an engineering organization, or a fund, share what you are working on. I respond within two business days.
or email pankaj@pankajrajan.com