Senior engineer, 13+ years in distributed systems. I build anti-financial-crime systems at a Nordic bank by day, and solo-ship agentic-AI products โ and win hackathons โ by night.
I care about AI that explains its reasoning, leaves an audit trail, and actually works in production.
- ๐ฆ Software Engineer @ Swedbank โ anti-financial crime & AML
- ๐ค Building agentic AI on Azure AI Foundry, A2A, and MCP
- ๐งฎ Researching GPU & quantum compute โ q1729, the quantum taxicab
๐ก๏ธ ARGUSMulti-agent compliance intelligence Microsoft Agents League โ AI Skills Fest 2026 ยท Reasoning Agents track โ
problem> Manual KYC/AML review doesn't scale, and unaudited AI decisions don't survive a regulator's audit. approach> Five specialist agents coordinated over A2A on Azure AI Foundry. Every finding is cited via Foundry IQ. Full audit trail. All decisions are explainable, reproducible, and regulatory-proof. impact> 100% auditable, citation-grounded regulatory lookups replacing manual screenings.
|
๐ฐ BASTIONA governed institutional-agent fleet for continuous access review All Things Agentic Hackathon 2026 ยท Fortified Enterprise Fleet track โ
problem> Access review is quarterly work performed on continuously changing permissions. Automating the scan isn't enough โ an institutional agent must remember prior human decisions, survive asynchronous retries, prove why it acted, and remain unable to turn suspicious input into a privileged write. approach> Read-only IAM review against the GCP project that runs it, including its own service identities. Deterministic code detects and scores findings; Gemini explains and routes already-minimized risk. Three institutional agents, one durable investigation identity โ no raw IAM binding crosses the model or human-notification boundary. impact> Humans receive counts and allowlisted categories, never bindings. 161 tests at 100% statement and branch coverage.
|
๐ก DRIFTGPU & AI infrastructure release intelligence OpenAI Build Week ยท Devpost โ
problem> Raw changelogs are noisy, unstructured, and full of false positives. Teams miss critical AI infrastructure updates. approach> High-precision release aggregation. Raw data โ dependency checks โ bounded, technical summaries. Built with FastAPI + pgvector semantic deduplication. impact> Converts raw, noisy changelogs into actionable release intelligence.
Code โ ยท Live app โ ยท API docs โ ยท Demo video โ ยท Devpost โ |
๐ง CONTINUUMDurable incident memory for cold-started agents CockroachDB ร AWS Hackathon 2026 โ Build with Agentic Memory โ
problem> Cold-started agents lose execution state. Multi-step workflows restart from zero, wasting compute and losing context. approach> Distributed checkpoint engine on CockroachDB. Restores 100% of execution state without pipeline restart. Mission-critical runtime guarantees. impact> Restores full execution state across distributed orchestration.
Code โ ยท Live app โ ยท Demo video โ ยท Devpost โ |
๐ BANKERS' WRAPPEDYear-in-review intelligence for financial workflows Backblaze Generative Media ยท Devpost โ
problem> Data-centric fintech teams want year-in-review insights. Existing tools are generic, not built for financial workflows. approach> Spotify Wrapped but for banking. Extracts transaction intelligence, generates insights, creates shareable year-end summaries.
Code โ ยท Live app โ ยท API docs โ ยท Demo video โ ยท Devpost โ |
-FFB000?style=flat-square){kind=icon}
-
ARGUS: Compliance Infrastructure That Believes Financial Access Is a Human Right โ techcommunity.microsoft.com ยท Guest post (July 2026)
Microsoft published my full write-up on the Educator Developer Blog, including how ARGUS coordinates five agents over A2A with citation-grounded risk scoring.
Lee Stott ยท Microsoft in #agentsleague (theme-aware image)
๐งฉ Agentic AI & Enterprise Intelligence I design AI systems with Azure AI Foundry, Foundry IQ, multi-agent orchestration, Agent-to-Agent (A2A) communication, and RAG with hybrid search โ built to be explainable, grounded, and production-ready.
๐๏ธ Distributed Systems & Backend Architecture 13+ years building scalable platforms with Java (Spring Boot, Quarkus) and Python (FastAPI) โ microservices, NoSQL, event-driven systems, and hybrid cloud across AWS, Azure, and OpenShift.
โ๏ธ AI Infrastructure, Performance & Compute I work at the infrastructure layer behind modern AI โ GPU computing, NVIDIA CUDA, model serving, vector search, and performance engineering.
๐ก Reliability, Observability & Platform Engineering I build resilient systems with Azure Functions, Service Bus, OpenTelemetry, Application Insights, and KQL โ telemetry pipelines that hold up to enterprise-grade reliability and governance.
Languages & frameworks
Agentic AI & LLM
Cloud & infrastructure
AI infra, GPU & observability
2025 โ Present ยท Software Engineer ยท Swedbank โ Stockholm, Sweden Anti-financial crime ยท AML platforms ยท 95%+ test coverage across unified multi-module architecture
2021 โ 2025 ยท Senior Java Developer ยท Viaplay Group โ Stockholm, Sweden Media & streaming on AWS + Kubernetes ยท ~30% perf gains ยท ~40% delivery-speed acceleration
MarโSep 2021 ยท Software Developer ยท Expleo Technology Nordic โ Gothenburg, Sweden Domain-driven microservices ยท ~50% faster onboarding via docs & workflow diagrams
2012 โ 2021 ยท Senior Software Engineer ยท IBM โ Sydney & Pune Regulated banking APIs for Westpac ยท ~25% response-time gains ยท Jenkins/Bamboo modernization
- q1729 โ Ramanujan optimization via bare-metal GPU. CUDA-Q + cuQuantum + NVIDIA NIM Nemotron
- llm-qlab โ LLM quantization benchmarks on consumer GPUs. Speed, VRAM, accuracy trade-offs
- pythonic-algorithms-lab โ CPU vs GPU profiling with empirical Big-O analysis. CuPy + Numba CUDA
- iq-series โ Hands-on Microsoft IQ notebooks. Foundry IQ, Work IQ, Fabric IQ
ย
ย
- NVIDIA Certified Professional: Agentic AI โ NVIDIA
- Machine Learning and AI Foundations โ LinkedIn Learning Community
- IBM Machine Learning Essentials โ IBM
- Python for Data Science โ IBM
- Kubernetes Administration โ The Linux Foundation
Full certification list on LinkedIn โ
Building trustworthy AI systems that explain their reasoning, leave an audit trail, and actually work in production.
If that's the kind of problem you're working on โ I'd love to talk.
github.com/iarjunganesh ยท arjunganesh.dev