Kaushal Dave
Kaushal Dave
Head of AWS Applied AI Solutions · Sydney

Builds AI that decides. Leads from model to enterprise.

Head of AWS Applied AI Solutions across APJ and EMEA. Two decades building decision systems for supply chain, global retail pricing and delivery speed, and leading the architects, scientists and engineers who put them into production.

APPLIED AI · FORWARD DEPLOYED ENGINEERING · SYDNEY
Explore the work
01

What I do

I lead Applied AI from strategy to production for global enterprises: the product direction, the architecture, the teams and the operating model that turn decision systems into running operations.

The work spans three vantage points: econometric systems behind global retail pricing and delivery speed decisions at Amazon, agentic supply chain planning with Amazon Connect Decisions at AWS, and AI decision platforms for the Fortune 1000 at Aera Technology.

I stay close to both the code and the customer. I set direction, lead teams, run the technical decision, and build the missing pipelines, validators, benchmarks and grounded LLM applications when the product needs them.

PRODUCT STRATEGY · ARCHITECTURE · FIELD DEPLOYMENT · OPERATING SCALE

Leadership

Applied AI strategyProduct vision and PRFAQ Global org leadershipExecutive advisory P&L and delivery

Applied AI

Forward deployed engineeringAgentic workflows Supply chainPricing and economics Amazon BedrockInference infrastructure

Build

PythonSQLSpark KafkaREST and microservicesAWS
02

Experience

JAN 2024 TO PRESENT

Head of AWS Applied AI Solutions

Amazon Web Services · APJ & EMEA · Sydney
  • Lead Applied AI solutions and supply chain across APJ and EMEA. Originated and ran a portfolio of enterprise PoCs from a standing start, generating $2M+ in new ARR.
  • Took the first customer in the world to production on Amazon Connect Decisions, then set a repeatable 14 day cycle from data received to first viable results that turned one off wins into a pipeline.
  • Building the APJ team for Amazon Connect Decisions, starting with a Senior Solutions Architect hire across India, Japan, Korea, Australia and Singapore.
  • Hands on across Bedrock, SageMaker, Glue and Athena. I run the technical deep dives, own my demos, and build the data onboarding pipelines, validators and visualizations a maturing service needs, so customers see the whole solution.
NOV 2022 TO DEC 2023

Head of Product, Economic Decision Science

Amazon, Economics and Decision Science (EDS Tech) · London
  • Owned strategy and roadmap for large scale econometric and ML systems behind global retail pricing and delivery speed decisions.
  • Directed engineering investment, non functional requirements and technical trade offs for systems influencing multi billion dollar outcomes.
  • Ran the operating mechanisms: product roadmaps, investment plans, QBRs and MBRs across scientists, engineers and product managers.
JUN 2017 TO NOV 2022

General Manager, Solutions & Product

Aera Technology · Singapore
  • Founding member reporting to the CEO. Built and led a global org from 4 to 130+ across solutions, engineering and data science.
  • Owned architecture and delivery for Fortune 1000 deployments with SAP and Oracle integration, real time pipelines and AI recommendations.
  • Operated platforms making 500M+ decisions a day at 99.9% uptime, influencing $2B+ in business outcomes.
  • Opened 3 greenfield offices, presented to the board quarterly and pitched investors directly.
NOV 2014 TO JUN 2017

Global Head of Enterprise Cloud Solutions, Supply Chain Apps

Anaplan · San Francisco
  • Built the supply chain line of business from zero to 62 customers and $100M+ ARR, while leading delivery of 100+ enterprise planning applications.
2007 TO 2014

Product and Strategy Leadership, earlier roles

E2Open · KLA · Siemens (Mentor Graphics)
  • Product strategy for cloud planning (E2Open); simulation and design software for 8 of the top 10 chip manufacturers (KLA); global ERP integration programs (Siemens).
03

Live work sample

Live · grounded in a private skill store

Most resumes describe the work. This one runs it.

The agent below runs live on a managed LLM, grounded in a private skill file and served through a serverless API. Ask about leadership scope, product strategy, technical depth or the systems I have built. The answers come from the record, not a fixed script.

Managed LLMServerless API Private skill storeRAG grounding Least privilege IAM
resume agent · live
Hi, I'm Kaushal's resume agent, running live on a grounded LLM. Ask me anything, or tap a question below.
04

Try the work

INFERENTIA2 · p50 latency
NVIDIA A10G · p50 latency
Measured · audited

Inference benchmark audit

120,000 requests on Inferentia2 and A10G. The latency and reliability result held. The throughput and cost claim did not, so it was withdrawn.

Read what survived the audit
Live product · try it

OC Practice

A live consumer AI app: a Bedrock backed engine generates adaptive test questions with trap distractors, automatic scoring and progress dashboards.

Live product · explore it

MEPDraft

AI native building services engineering: a CAD and DXF geometry parser, a plumbing code rule engine and automated drawing generation.

05

Selected writing

Forward Deployed Engineering: from model to enterprise
Forward Deployed Engineering

From model to enterprise

A working model is one layer. Deployment is the method that makes it grounded, connected and repeatable.

9 minute read
Applied AI in production: what the operating model requires
Applied AI in production

The operating model required

What the organisation declares, what the system derives, and who owns the result.

8 minute read
120,000 inferences: the economics of purpose built AI infrastructure
Benchmark audit

The result that survived

The latency result held. The economics claim did not. The audit records both.

8 minute read
06

Leadership in practice

Set the direction. Stay close to the code. Put the system in front of customers. Then build the organisation that can repeat it at global scale.

Product to production

Owns the path from product strategy and architecture through operating mechanisms, field deployment and measurable business outcomes.

Code to C suite

Builds the pipelines, validators, benchmarks and grounded applications that fill product gaps, then carries the same work into executive decisions.

One market to global scale

Built and led across APJ, EMEA and the Americas: a global org from 4 to 130+, three greenfield offices and enterprise systems operating every day.

Sydney · APJ

Let's put AI
into production.