About
Engineering Manager at ANZ working on cloud-native platforms, developer experience, and bringing agentic AI into the software delivery lifecycle.
Dipjyoti Metia
Engineering Manager · Melbourne, Australia
I'm an engineering manager who still builds — happiest working on the platforms and tooling that make other teams faster. I work across microservices, event-driven systems, and cloud-native architecture — mostly in Golang and TypeScript — and I'm increasingly focused on putting agentic AI and LLMs to practical use across the delivery lifecycle: test generation, evaluation, and the human-in-the-loop guardrails that keep them safe in production.
What I believe
Agentic AI earns its place in the delivery lifecycle only when it's paired with evaluation and human-in-the-loop guardrails. The goal isn't automation for its own sake — it's giving engineering teams real leverage (test generation, scenario modelling, natural-language tooling) without trading away the reliability, observability, and quality that production systems depend on.
Professional Journey
As Engineering Manager for Enablement Engineering at ANZ, I lead the platforms and shared tooling underneath core digital banking services, with a focus on reliability, observability, and developer experience. The role runs from strategy through to hands-on engineering — and an increasing part of it is working out where agentic AI and LLM-assisted automation genuinely speed teams up without trading away safety or quality.
- Platform Engineering
- Quality Engineering
- Agentic AI
- Team Leadership
Technical Skills
- Golang
- gRPC
- Python
- TypeScript
- Docker
- Kubernetes
- Google Cloud
- AWS
See my full background and certifications on the résumé .
Connect your AI to this site
This site runs a Model Context Protocol (MCP) server — so instead of asking an AI to guess about my work, point your own assistant (Claude, Cursor, …) at it and query my writing, projects, and experience straight from the source. Or try it right here — this calls the live server:
// type a query and hit run — or tap a tool above
add it to your own client →
endpoint
https://dipjyoti.dev/mcp/
{
"mcpServers": {
"dipjyoti-dev": { "url": "https://dipjyoti.dev/mcp/" }
}
} Tools: search_writing, get_post, list_projects, get_experience.
Recent Writing
- genaiai-agents
Containing AI Agents at the OS Level: A Hands-On Look at Microsoft's MXC SDK
Microsoft's MXC SDK wraps model-generated agent code in policy-driven OS sandboxes — what it isolates, how to wire it up, and where the early preview falls short.
- llmobservability
Evaluating an LLM Agent Like Real Software: Observability and Evals with Langfuse
A vibe-check isn't a test. How to trace, score, and gate an LLM agent with Langfuse — and the silent escalation regression evals catch that a demo never would.
- genaillamaindex
Fine-tuning Generative AI Models: A Practical Guide with LlamaIndex and Modern Frameworks
Explore the latest techniques for fine-tuning generative AI models using LlamaIndex, Hugging Face Transformers, and parameter-efficient methods like LoRA and QLoRA. This guide covers everything from data preparation to deployment, ensuring your AI applications are contextually relevant and efficient.