Staff Software Engineer
I build and lead the design of distributed, well-tested backend systems — and the infrastructure that makes enterprise AI safe to ship. Over sixteen years across C#/.NET, Azure, and Kubernetes, most recently building LLM evaluation and observability platforms.
01 — About
I'm a product-minded Staff Software Engineer with 16+ years of experience, working primarily in C#/.NET on Azure. My core is distributed systems, microservices, and domain-driven design — building services that are well-tested, maintainable, and operationally owned end to end.
Most recently I've focused on enterprise AI infrastructure: declarative LLM evaluation frameworks for safe model-swapping, and observability over full LLM interactions — the backbone of a triage → evaluation → fix loop used across multiple AI-agent teams. I hold an MSc in IT with an NLP focus, and I use Claude Code and GitHub Copilot daily.
I work across the stack when it helps the team ship, and I lean on strong fundamentals — databases, testing, and clear architecture — over chasing tools. I'm based in West London.
02 — Selected work
Observability over full LLM interactions — user turns, system prompts, tool calls, context — as the backbone of a triage → evaluation → fix loop across ~30 engineers and six AI-agent teams, in partnership with InfoSec. Product managers and engineers reproduce and share real interactions ~4–5× faster than the old copy-paste-and-repro loop.
A declarative framework with 20+ evaluators for non-deterministic testing and safe LLM model-swapping — every fix must pass all prior evaluations before it ships. Human-readable scenarios are decoupled from config and evaluators, so turning a bug into a new eval is trivial — driving fast, org-wide adoption (onboarding dropped from ~a week to ~3 hours).
Led a zero-downtime consolidation of core AI features into a single unified service across a 40+ engineer codebase — cutting cognitive load so teams shipped faster.
In an empowered product trio, I caught an off-roadmap customer pain during discovery and shipped a database-style visualisation as a ~2-day experiment with one frontend engineer — one of the year’s most successful features, shown by the CPO at all-hands.
03 — Open source
Pulled from GitHub at build time, so the state below reflects each repo as of the last deploy.
This is a repository to keep my useful Claude code skills
Building in public — an AI agent at scaffold stage, developed with a strict vertical-slice method.
This code is to let users select a simple text in Windows and then it reads it out for them. They can control the spead as well.
04 — Experience
C#/.NET, Azure, AKS, DDD, event-driven microservices; AI infrastructure and technical leadership.
Client-facing delivery across multi-country sites; C#/.NET, Azure, SQL Server.
Scaled the platform from 500k to 900k+ users; proposed simpler, often better implementations.
Built in-house and client-facing software, and coached junior developers. .NET Framework, SQL Server, SharePoint.
Credentials & honours
05 — Recommendations
During incidents, he was someone you could genuinely count on: calm, sharp, and focused on solving the problem in the right way. His expertise in performance, scalability, and monitoring of complex systems, as well as his deep database knowledge repeatedly helped resolve difficult issues and close gaps in situations where others were stuck.
Dan Antal Staff Engineer · ~7 years at Templafy He has a rare ability to move effortlessly between high-level architecture and low-level implementation, quickly understanding how components fit together and identifying the right trade-offs. This made him a key contributor to our distributed architecture, where his impact was both significant and lasting.
René Gundersen CTO · managed Ashkan at Templafy In refinements he was known for his drawings that would clarify the solutions we were refining and the constraints of the software… It was the same Ashkan I saw mentor and develop engineers, and turn each one of them into confident problem solvers.
Flemming Humlum Søgaard Product Manager · Templafy What stood out was his ability to bring structure to complex areas like AI evaluation and data flows, translating risk considerations into scalable, production-ready systems.
Silvia Belloni InfoSec & IT Governance Lead · Templafy Ashkan has a rare gift for bridging the gap between high-level business strategy and low-level system design. He consistently acts as a force multiplier for his organization. Any top-tier engineering team would be fortunate to have his vision and technical depth.
Hossein Margani DevOps Engineer · WTW