Hey, I'm Inthu 👋

AI Evangelist & Strategic Advisor at AWS

Inthuson Anandakumar

Half my time is spent advising C-suite leaders at insurers, banks, and capital markets firms on cloud and AI strategy. The other half is spent hands-on with AI tools, writing code and building — I like really knowing what I'm talking about when I'm advising on it.

The AI thing started early

My first encounter with AI was in 2016, doing research on Probabilistic Latent Semantic Analysis at the University of Leicester — a precursor to the Transformer architecture that powers today's LLMs. Looking back, I probably should've stuck with research. But when I graduated, I couldn't pass up the opportunity to go into defence — and that detour through national security, government, and cloud gave me something I wouldn't trade: I know what it takes to ship AI in heavily regulated environments where you can't just "move fast and break things."

Before AWS

I was at Leonardo and then 6point6 (since acquired by Accenture), working across defence, national security, and government. I can't talk about most of those projects — but the work touches people's lives every single day, even if no one knows about it. I love seeing it in action going about my day to day.

At AWS

I'm a trusted technical advisor and strategic partner to some of the largest financial institutions in the world, helping them define and execute their first major AI transformations — the kind of projects that fundamentally change how these organisations compete.

Earlier at AWS I built a service lifecycle management tool that saved over 194,000 manual man-hours in 6 months. It worked well enough that it got integrated into AWS Health and adopted across the global organisation. I also worked directly with the Bank of England and FCA on incident management frameworks, led crisis response for critical financial services incidents, and spoke at AWS Summits across London, Paris, and Dubai.

Open Source

Contributions to public repos — mostly AWS-adjacent stuff, serverless patterns and tooling fixes.

1.8k

Authored a new pattern: accessing Bedrock via API Gateway with Cognito auth, domain restriction, and rate limiting. Full CDK Python app — 19 files, merged.

CDKPythonBedrockCognitoAPI Gateway
View PR →

Amazon Transcribe Post Call Analytics

aws-samples/amazon-transcribe-post-call-analytics
107

Fixed the build script so it actually works on Apple Silicon. The publish.sh was assuming x86_64 — set the arch explicitly so M-series Macs can build locally.

BashDockerARM64
View PR →

Most of my code lives in private repos. GitHub profile →

These days

I mostly use AI tools to build and ship — it's changed how I work completely. I invest in and help a handful of early-stage startups. And I'm passionate about teaching — I started back in sixth form running coding workshops for kids, and now I give talks to university students and at schools. Because the kid who didn't know anyone in tech could've used someone like that.