Product, AI & Data Executive

London Toronto Seattle

Anuj Batra

Turning AI ambition into AI advantage.
AI Product Executive · Builder · Fractional Leader

AI Product Executive, builder, and technical leader with 25+ years across Meta, Microsoft, IBM, Loblaw, and KiloVision AI. My primary discipline is product: defining what to build, why it matters, and how to bring it to market. I combine executive product leadership with deep technical fluency, leading AI products, platforms, and enterprise transformations for CIOs, CEOs, and boards while remaining hands-on in architecting and building production AI systems from first principles.

Get in Touch → View Case Studies

Experience across

Meta IBM Microsoft Loblaw Element AI CI&T SAP Siebel

Currently building

KiloVision AI — veterinary radiology AI with computer vision, LLM-powered clinical reporting, and end-to-end workflow integration View case study →

$2.3B+

Attributed revenue impact

Revenue attributed to programs led at Meta

25+

Years in product, data & technology

Director to C-suite across 3 continents

300+

Largest team led

From focused teams to 300+ at CI&T; 120+ at Loblaw

Industries delivered across

Deep domain expertise in retail/commerce and healthcare AI

Technical fluency

Engages directly in architecture reviews, model evaluations, and technical design sessions. Not as an observer, but as a participant with an informed point of view.

As CPTO at KiloVision AI, led the engineering team building the full AI stack while remaining hands-on in core architecture decisions. Designed evaluation and ML experimentation frameworks at Meta and Loblaw alongside engineering teams. Delivered classical ML systems across healthcare (IM Knowledge, clinical risk stratification), supply chain (Element AI, RNN demand forecasting), and retail (Loblaw, recommendation and personalization at 30M+ user scale).

Technical areas

Large Language Models RAG Architectures Agentic AI Computer Vision Prompt Engineering ML Experimentation & Evaluation Recommendation Systems A/B Experimentation at Scale Data Platform Architecture Responsible AI & Governance Observability & Drift Detection Production ML Systems

Where applied

⚡

LLM & agentic AI

Built at KiloVision AI (RAG, clinical report generation); LLM-powered agent at Meta (dynamic ad creative personalization to maximize relevance and response rates)

🔬

Experimentation platforms

Designed Meta's evaluation framework; built Loblaw's ML and A/B platform adopted across business units

🏗

Data platforms

Architected data infrastructure at 30M+ user scale (Loblaw) and ad auction scale (Meta)

📐

Classical ML

RNN demand forecasting (Element AI), clinical risk stratification (IM Knowledge), personalization & collaborative filtering (Loblaw)

What I bring

From strategic direction to hands-on delivery.

Effective across the full range: executive leadership of complex, multi-stakeholder AI and data programs at enterprise scale, and the hands-on execution of a founder building production systems from first principles.

🏗

Product builder

As CPTO, led the engineering team that built KiloVision AI's stack (LLM pipeline, RAG architecture, computer vision, clinical guardrails) while staying hands-on in core architecture. At Meta, led the technical development of an agentic AI system at auction scale.

📊

Enterprise AI leadership

Led complex, multi-stakeholder AI and data programs at Meta, Loblaw, and CI&T, leading teams of up to 300+ across data engineering, data science, ML engineering, and software engineering, owning commercial outcomes alongside executive sponsors.

🔁

Data and AI depth, end to end

Full-stack operational experience: data strategy and platform architecture through model experimentation, deployment, and observability. Built the data infrastructure at Loblaw, rebuilt the signal pipeline at Meta, led the GenAI stack build at KiloVision AI.

Case Studies

Selected case studies.

Four high-impact case studies across AdTech, enterprise data and AI transformation, healthcare, and enterprise product platforms. Each demonstrates quantified, verifiable business outcomes and direct executive accountability.

AdTech · Agentic AI →

Meta · Ads Monetization

AI-Driven Ad Platform: Autonomous Experimentation & Signal Growth

Two concurrent programs: an agentic AI system for autonomous creative experimentation, and a Signal Growth initiative to recover and expand data signal for ad targeting, both at platform scale across millions of daily auctions.

$2.3B+ revenue impact 18%+ ad relevance lift

Agentic AISignal GrowthData strategyAdTech

Healthcare AI · Founder →

KiloVision AI · Founder & CPTO

Veterinary Radiology AI Platform & Clinical LLM

Founded and built KiloVision AI, applying computer vision, NLP, and large language models to veterinary radiology. Full CPTO ownership across product strategy, ML architecture, GenAI pipeline, and commercial outcomes.

97% diagnostic accuracy 0 → production in <12 months

GenAIComputer VisionRAGClinical AI

Retail · Enterprise Data & AI →

Loblaw Digital · VP Product

Enterprise Data & AI Transformation — Canada's Largest Retailer

Led enterprise-wide data and AI transformation at Loblaw: 30M+ users, 120+ person org, driving 37% eCommerce growth and 17%+ AOV through personalization ML, data platform modernisation, and medication adherence AI.

37% eCommerce growth 17%+ AOV lift 120+ org built

Data platformPersonalization MLA/B at scaleHealth AI

AI Products · Enterprise →

CI&T · Head of Data & Analytics, AI Products & Platforms

Enterprise AI Commerce Platform & Developer SDK

Built and led CI&T's AI products practice from the ground up: a 300+ person organization delivering production AI commerce programs and a developer SDK for Fortune 500 clients across retail and life sciences.

$130M+ revenue generated 300+ org built and led

FSRetailHealthcare0→1

View all case studies →

Services

Three ways to engage.

From embedded product leadership to AI strategy and program leadership, through to hands-on execution. Each scoped to the specific problem, the team, and the outcome required.

01 — PRODUCT

AI Product Management

Fixed-term, embedded: 0→1 through launch and GTM

A senior PM mandate on contract, embedded with the team to own the roadmap, the engineering relationship, and the launch.

View details →

02 — STRATEGY & PROGRAM LEADERSHIP

AI Strategy, Governance & Program Leadership

For boards, CIOs, and organizations of any size

AI strategy, governance, change management, and day-to-day program execution for organizations introducing AI into an existing business, or scaling a dedicated product function.

View details →

03 — EXECUTION

Enterprise AI Delivery & Execution

Enterprise AI programs for CIOs, CEOs & CTOs

Embedded leadership for complex AI and data programs, from strategic alignment with CIOs, CEOs, and boards through to hands-on technical execution across enterprise and startup contexts.

View details →

A perspective on AI product leadership

The hardest part of building AI products is not the model. It is everything around it.

Most AI programs that underdeliver do not fail because the model was wrong. They fail because the problem was framed too narrowly, the data strategy was an afterthought, the evaluation criteria were unclear, the production system was never designed for the operational reality of a learning system, or, most often overlooked, user adoption was treated as a launch activity rather than a design constraint.

A model that is never trusted is a model that was never shipped.

The 12-Stage AI Delivery Framework below is the operating model built and refined across six AI programs, from rebuilding Meta's ad signal architecture after platform-wide signal loss, to founding and scaling KiloVision AI's clinical diagnostic pipeline, to leading enterprise AI transformation at Canada's largest retailer. It is not a generic industry reference. Each stage reflects a specific failure mode encountered, and fixed, in production.

Where value is actually created

The organizations that extract consistent value from AI share a common pattern: they invest as heavily in the problem definition, data strategy, evaluation design, and governance infrastructure as they do in the models themselves. The model is one component. The system and the product decisions surrounding it determine the outcome.

What good leadership looks like

It means asking harder questions earlier: Is this the right problem? Do we have the data to solve it? How will we know if the model is working, or failing quietly? Who is accountable when it drifts? These are product and leadership questions as much as they are technical ones. Getting them right from the outset separates programs that scale from those that stall.

01 — Opportunity

Identification & Business Case

Determining where AI creates genuine value, and where it does not. Customer research, market analysis, AI feasibility assessment, ROI estimation, and success metric definition across both business and ML dimensions.

Problem framing AI feasibility KPI design Business case

02 — Discovery

Requirements & Constraints

User research, journey mapping, stakeholder alignment, regulatory assessment, and technical feasibility, including decisions on prediction vs. generation, human-in-the-loop, explainability, latency, and cost constraints.

PRD User research Regulatory review Non-functional requirements

03 — Data Strategy

Data Acquisition & Governance

Identifying data sources, collection, labeling, annotation, privacy review, and data contracts, with explicit consideration of bias, coverage, freshness, completeness, and drift risk.

Data pipeline Labeling Data governance Drift risk

04 — Solution Design

Architecture & Model Strategy

Selecting the right approach (rules engine, classical ML, deep learning, LLM, RAG, agentic AI, or hybrid) and designing the end-to-end architecture: orchestration, model layer, tooling, vector DB, and enterprise system integration. Lesson from Meta: when one layer of a system breaks, treat the redesign as a system problem, not a single-component fix.

LLM / RAG Agentic AI Architecture design Model strategy

05 — Experimentation

Prototyping & Model Evaluation

Feature engineering, prompt engineering, baseline modeling, model comparison, fine-tuning, and systematic evaluation across accuracy, hallucination rate, latency, cost, and task-specific metrics (BLEU, ROUGE, F1).

Prompt engineering Fine-tuning Evaluation Hallucination rate

06 — MVP Development

Build & Integration

API development, prompt orchestration, guardrails, feedback mechanisms, confidence scoring, fallback logic, and human review workflows, building a production-ready system, not just a proof of concept.

Guardrails Orchestration Feedback loops Fallback logic

07 — Validation

Business, User & AI Evaluation

Evaluating across three dimensions simultaneously: business outcomes (adoption, revenue, cost savings), user experience (satisfaction, trust, task completion), and AI performance (accuracy, hallucinations, bias, robustness), via alpha, beta, A/B, and shadow mode. Lesson from IM Knowledge Group: a model nurses do not trust gets ignored regardless of accuracy. Adoption is earned incrementally, not asserted at go-live.

A/B testing Shadow mode Go/No-Go Adoption design

08 — Deployment

Production & Infrastructure

CI/CD, model registry, containerization, canary rollout, feature flags, autoscaling, and production infrastructure (Kubernetes, GPUs, vector DB, feature store) with monitoring from day one.

CI/CD Canary rollout Model registry MLOps

09 — Observability

Monitoring & Performance Management

Continuous monitoring of latency, cost, accuracy, data drift, hallucinations, prompt failures, token usage, and business metrics, because AI system performance degrades silently in ways traditional software does not.

Drift detection Latency Token cost Operational dashboards

10 — Continuous Learning

Feedback, Retraining & Improvement

Collecting feedback, retraining models, updating prompts, refreshing knowledge bases, and improving orchestration, operating the closed-loop cycle of feedback, data, training, evaluation, deployment, and monitoring that keeps AI systems improving.

Retraining RLHF Knowledge refresh Closed-loop

11 — Responsible AI

Governance Throughout the Lifecycle

Privacy, security, compliance, fairness, explainability, auditability, model versioning, human oversight, and risk management. Not as a gate at the end, but as a continuous discipline embedded across every stage of the program. Lesson from IM Knowledge Group: in regulated environments, the audit infrastructure is not optional. It is the legal condition under which the system is permitted to operate at all.

AI governance Model cards Data lineage Human oversight

12 — Scaling

Growth, Optimization & Platformization

After product-market fit: multi-region deployment, cost optimization, model routing, multi-model strategy, personalization, enterprise integrations, and platformization, with unit economics, reliability, and AI cost per transaction as the governing KPIs.

Model routing Cost optimization Platformization Unit economics

Services mapped to this lifecycle

Product — stages 1–6: opportunity through MVP | Strategy & Governance — stage 11, runs throughout | Execution — stages 7–12: deployment through scale

Start a conversation.

If you are building an ambitious AI or data product, leading a complex program, or exploring a senior leadership hire, reach out directly.

Get in Touch →

About

AI Product Executive. Enterprise AI Leader. Builder.

25+ years building and leading AI, data, and technology organizations across Meta, IBM, Microsoft, Loblaw, CI&T, Element AI, IM Knowledge Group, and KiloVision AI, across London, Toronto, and Seattle.

Career scale

$2.3B+ revenue attributed to programs led
300+ cross-functional teams led (CI&T)
120+ org built and led (Loblaw)
30M+ user platforms in production

AI track record

GenAI + RAG pipeline built (KiloVision)
Agentic AI at auction scale (Meta)
Clinical AI in regulated environments (IM KG)
$150M+ AI forecasting impact (Element AI)

Operates across the full spectrum: setting AI and data strategy at board level, leading teams of 300+ within enterprise organizations, and leading engineering teams as a hands-on founder. At KiloVision AI, led the engineering team building the LLM pipeline, RAG system, and computer vision stack as CPTO, while staying directly involved in core architecture decisions. At Meta, designed the agentic AI evaluation framework. At Loblaw, built the ML experimentation platform alongside engineering teams.

Open to senior leadership roles across AI product, data, and technology: UK, Europe, US, and Canada.

Career

25 years of leadership across product, platform, and AI.

2023 — present

Founder & CPTO (Chief Product & Technology Officer)

KiloVision AI

Founded and leads a veterinary radiology AI company, with full ownership of product strategy, ML architecture, engineering delivery, and commercial outcomes. Built from concept through to production.

2022 — 2023

Technical Product Leader, Ads Monetization

Broad capability. Concentrated depth.

The value of genuine cross-functional fluency across strategic, technical, and commercial disciplines is the ability to operate credibly at every layer of a complex organization, without requiring translation between them.

AI Product Strategy Generative AI & LLMs Full AI Lifecycle Leadership AdTech & Auction Systems RAG & Agentic AI ML Model Governance Data Platform & Architecture Computer Vision Recommendation Systems A/B Experimentation at Scale Data Strategy & Governance AI Governance & Risk Technical Program Management P&L Ownership Cross-functional Leadership Executive Communication API & Developer Ecosystems Healthcare AI Commerce Intelligence Organizational Design Clinical AI & Healthcare

Services

Product, data and AI leadership, from strategy to hands-on execution.

Three ways to engage, each scoped as a defined contract or fractional commitment rather than an open-ended retainer: embedded product leadership, AI strategy and program leadership, and hands-on execution. The work spans leading small focused teams through to teams of 300+.

How the work gets done

Commercial & technical architecture, together

Technical initiatives are directed, not just sponsored, including the data contracts, evaluation metrics, and compliance layers underneath the model, and tied directly to the commercial problem they're solving.

Executive and board-level alignment

Converting ambition into a structured, prioritized roadmap with CEOs, CIOs, and boards, with capital allocation tied to verifiable outcomes rather than technology for its own sake.

Ownership of the hard trade-offs

Latency against accuracy, speed against data quality, scope against deadline: these calls get made directly, including phasing a rollout against executive pressure when the data isn't ready.

Adoption as the design brief

Privacy, interface design, and fallback logic are treated as core product constraints from day one, not compliance add-ons. The goal is sustained use, not just deployment.

$130M+ AI products practice (CI&T) $2.3B+ revenue recovery program (Meta) 37% eCommerce growth, 120+ org (Loblaw) 0→1 healthcare AI, founding to production (KiloVision)

01 — PRODUCT

AI Product Management & Leadership

Fixed-term, embedded product leadership: 0→1 through launch and GTM

A senior PM mandate, taken on contract. Embedded as the product leader for a defined engagement, owning the roadmap, the engineering relationship, and the launch, with AI woven in wherever it's the right tool rather than treated as the point of the exercise. This is product leadership work, not advisory work from the outside.

Opportunity identification, AI feasibility assessment, and business case development
User research, journey mapping, and requirements definition (PRD, user stories)
Product vision, roadmap design, and prioritization frameworks
AI solution design: model strategy, architecture, and build-vs-buy evaluation, where AI is the right tool for the problem
Experimentation design, evaluation frameworks, and go/no-go validation
Day-to-day engineering execution: sprint cadence, backlog ownership, and delivery accountability
Go-to-market strategy: pricing model design, packaging, positioning, and launch execution
Channel strategy, partner and ecosystem development, and enterprise sales enablement
Adoption measurement, commercial performance tracking, and 1→10 scaling

Engagement: fixed-term contract, typically 3 to 12 months, embedded with the team

Relevant for: CPOs and CEOs needing a senior product leader on contract to drive a specific 0→1 or scale mandate

02 — STRATEGY & PROGRAM LEADERSHIP

AI Strategy, Governance & Program Leadership

Driving AI direction and execution, including where product is not the primary function

Senior leadership for organizations setting AI direction and then actually running it: governance, change management, stakeholder alignment, and day-to-day program execution. Built for organizations introducing AI into an existing business, as well as those running a dedicated internal product function.

Enterprise AI strategy, portfolio prioritization, and innovation program leadership
AI roadmap design across product lines, business units, and functions
ML model governance, risk frameworks, and audit design
Responsible AI: fairness, explainability, bias evaluation, compliance
Change management and stakeholder alignment across business, technology, and regulatory functions
User and organizational adoption strategy, not just technical rollout
Day-to-day technical program direction: cross-functional delivery, risk tracking, and executive reporting
AI organizational design, team structure, and capability building
Investor and board-level AI narrative and due diligence support

Engagement: fixed-term contract or fractional retainer, typically 3 to 9 months

Relevant for: Boards, CEOs, CIOs, and COOs at enterprises, multinationals, and regulated institutions running an AI strategy or innovation mandate, at any company size

03 — EXECUTION

Program Leadership & Hands-on Execution

AI and data programs across organizations of every size

Embedded product and program leadership for organizations that need senior hands-on execution, working directly with the team rather than reviewing progress from a distance. Covers the full delivery scope: product management, cross-functional coordination, technical alignment, and stakeholder management across ambiguous, high-stakes programs. Effective leading teams at any scale, from a focused startup team through to teams of 300+.

Senior product management and roadmap execution within engineering teams
AI and data program leadership from concept through production and scale
Cross-functional coordination: data engineers, data scientists, ML engineers, software engineers, and business stakeholders
Architecture alignment, technical decision governance, and design review participation
Experimentation design, evaluation frameworks, and production validation
Data strategy, governance, and data quality program leadership
Stakeholder management, delivery risk identification, and program recovery

Engagement: fractional or embedded, typically 2 to 5 days per week

Relevant for: CTOs, Heads of Engineering, CPOs, and leadership teams running AI and data product programs

Domain depth

Retail & Commerce

Deep domain experience from Loblaw Digital (VP, Commerce Intelligence Platform, Canada’s largest retailer) and CI&T (AI product programs for Fortune 500 retail organizations), spanning personalization, loyalty ML, demand forecasting, recommendation systems, and digital commerce platform architecture at national scale.

Healthcare AI

Substantive healthcare AI experience from two distinct contexts: founding KiloVision AI (veterinary radiology, computer vision, LLM-powered clinical reporting) and leading the clinical AI platform at IM Knowledge Group (triage risk stratification, infectious outbreak detection, and COVID-19 vaccination prioritization) deployed in regulated HIPAA/GDPR environments. Governance, explainability, and clinical validation across both.

How I work

Embedded. Accountable for the outcome.

The model is simple: I join as an extension of the team and own the commercial outcome, not just the deliverable. Not an external voice commenting from the outside, and not scope without accountability. Whether leading a product workstream, running a program, or shaping AI strategy, the work carries the same standard: measurable results, tied to the business case that justified the engagement.

01 — Start

Initial conversation

A direct, focused discussion, typically 30 to 45 minutes, to understand the business problem, the commercial stakes, and what success looks like. No pitch deck, no credentials presentation. An honest assessment of fit from both sides.

02 — Scope

Scope definition

A clear, written scope of work with defined objectives, deliverables, and success metrics tied to commercial outcomes. Engagements run from four weeks to six months or longer, structured around results. Commercial terms are straightforward and agreed upfront.

03 — Execute

Embedded execution

From day one, operating as part of the team and owning delivery: in product reviews, architecture discussions, stakeholder sessions, and technical evaluations. This may mean leading a product workstream, running a program end to end, or providing the senior product or AI leadership a team needs to move faster, with direct accountability for the result.

→

A note for hiring teams

The same approach applies to permanent senior leadership roles. Whether the context is an engagement or a long-term position, the work is embedded, accountable, and focused on outcomes from day one.

Start with a conversation.

The most useful first step is always a direct discussion about the problem you are trying to solve, the team you have, and what the right outcome looks like.

Get in Touch →

Case Studies

Work that moved the needle, at platform, enterprise, and startup scale.

From a $2.3B+ revenue program at Meta to enterprise data transformation at Canada's largest retailer to founding and building an AI company from first principles. Each case study reflects direct ownership and specific, verified outcomes.

AdTech · Agentic AI →

Meta · Ads Monetization

AI-Driven Ad Platform: Autonomous Experimentation & Signal Growth

Two concurrent programs at Meta: an agentic AI system for autonomous creative experimentation, and a Signal Growth initiative expanding data signal for ad targeting, both operating at platform scale across millions of daily auctions.

$2.3B+ revenue impact 18%+ ad relevance lift

Agentic AISignal GrowthData strategyAdTechML systems

Healthcare AI · Founder →

KiloVision AI · Founder & CPTO

Veterinary Radiology AI Platform & Clinical LLM

Founded and built KiloVision AI, applying computer vision, NLP, and large language models to veterinary radiology interpretation and clinical workflow. Full CPTO ownership: product strategy, ML architecture, GenAI pipeline, and commercial outcomes.

97% diagnostic accuracy 85%+ workflow efficiency gain

GenAIRAGComputer VisionClinical AILLM

Retail · Enterprise Data & AI →

Loblaw Digital · VP Product

Enterprise Data & AI Transformation — Canada's Largest Retailer

37% eCommerce growth 17%+ AOV lift 30M+ users

Data platformPersonalization MLA/B at scaleHealth AI

AI Products · Enterprise →

CI&T · Head of Data & Analytics, AI Products & Platforms

Enterprise AI Commerce Platform & Developer SDK

Built and led AI commerce platform and ML-as-a-service SDK for Fortune 500 clients across retail and life sciences. Created a licensing revenue model. Led 300+ globally.

$130M+ revenue generated 300+ global team

Commerce AIDeveloper SDKML-as-a-ServiceFinServ · Life Sciences

Supply Chain · Demand Forecasting →

Element AI · Director, AI Product Management

AI Demand Forecasting — Global Luxury Beauty Brand

Led production-grade AI forecasting system from problem framing through deployment: RNN time series modeling, multi-source data integration, and operational workflow redesign across B2C, B2B, and manufacturing channels.

90%+ forecast accuracy $150M+ impact 30%+ OOS reduction

RNN · TensorFlowSupply chainGCPRetail AI

Clinical AI · Healthcare →

IM Knowledge Group · VP, AI/ML Products

Clinical AI Platform — Triage, Outbreak Detection & Public Health

Led three-track clinical AI platform — triage risk stratification, infectious outbreak detection, and COVID-19 vaccination prioritization — deployed across regulated healthcare environments. Vaccination model adopted by public health authorities.

90%+ predictive accuracy 30% adverse outcome reduction

Clinical AIHIPAA/GDPRPublic healthRegulated AI

Anuj Batra

From strategic direction to hands-on delivery.

Selected case studies.

Three ways to engage.

The hardest part of building AI products is not the model. It is everything around it.

Start a conversation.

AI Product Executive. Enterprise AI Leader. Builder.

25 years of leadership across product, platform, and AI.

Broad capability. Concentrated depth.

Product, data and AI leadership, from strategy to hands-on execution.

AI Product Management & Leadership

AI Strategy, Governance & Program Leadership

Program Leadership & Hands-on Execution

Embedded. Accountable for the outcome.

Start with a conversation.

Work that moved the needle, at platform, enterprise, and startup scale.

Get in touch.

Send a message

Almost there