My Engineering Philosophy
I believe an engineer is a problem solver who should be capable of addressing any type of challenge they encounter. My approach is centered on deconstructing the “why” and “what” of every system to understand it inside out. By identifying architectural patterns, I can solve complex problems more efficiently and adapt to new technologies quickly.
Who I Am
I’m an engineer with 1.5 years of experience building AI and full-stack systems. I specialize in agentic applied AI: conversational voice AI, coding agents, multi-agent systems, research and workflow automation agents, evaluation pipelines, and the infrastructure that makes them reliable, measurable and fast. I also work across full-stack engineering, DevOps and security.
I’m a Software Engineer at The Product Highway, and before that I worked at For Real and Sopra Banking Software. Some of what I’ve done:
- Cut voice-AI latency by about 45%, saving about 1.75 s per spoken message for US users on real traffic, with Redis tool caching on top.
- Fine-tuned a model to predict tool calls, served on GPU at 28 ms: exact tool selection rose from 0/50 to 18/50 and LLM calls per step fell 42%.
- Built a voice-and-text shopping assistant into a live storefront, calling server tools and on-page actions from live page and cart context.
- Architected a marketplace’s order lifecycle in MedusaJS, from exchanges and partial cancellations to fulfilment, with sagas, idempotency and retries.
- Handled DevOps end to end, designing CI/CD pipelines and deployments with GitLab CI, Docker, Helm, Kubernetes and AWS, and a Go CLI that automates a platform’s whole DevOps flow.
- Owned Open Banking APIs end to end, from design decisions through implementation to QA: 10+ Spring Boot endpoints for the Berlin Group AIS and PIS standards, reaching full QWAC/QSEAL regulatory compliance.
- Worked across the security stack, from cryptography and HSM-managed keys to JWE encryption, JWT, OAuth 2.0 with mTLS, Keycloak and TLS.
- Worked forward-deployed with European banking clients, taking on the bugs raised in client meetings, debugging them end to end, and shipping quick fixes and hotfixes at delivery time.
I’m a problem solver and a passionate engineer. I like taking a system apart to understand why it works, learning new tools by building real things with them, and following a problem from the infrastructure all the way up to the product. The notes below are the working record: AI, system design, security, Kubernetes, Docker and more.
My Toolkit
| Category | Technologies |
|---|---|
| AI and Agents | Agentic workflows, tool calling, MCP servers, agent orchestration, agent memory, context engineering, human-in-the-loop and guardrails, evals and LLM-as-judge, RAG, vector databases, LangChain, Claude Code, Codex, conversational voice AI (streaming STT and TTS pipelines, latency tuning), fine-tuning |
| Programming Languages | Python, TypeScript, Java, Go, Kotlin, JavaScript, SQL, C++, C |
| Backend | FastAPI, Spring Boot, Spring Security, Gin, NestJS, MedusaJS, RESTful APIs, gRPC, GraphQL, WebSockets, Kafka, BullMQ, saga pattern, Postgres, MongoDB, Redis |
| Frontend | React, Next.js, React Native (CLI and Expo), Electron, TypeScript, Zustand, Redux, TanStack Query, HTML5, CSS3, Tailwind CSS, Bootstrap, Radix UI, Vite, Reanimated, React Navigation, Jest, Vitest, Figma-to-code, Stripe, PayPal and Klarna payment integration, Mixpanel and Adjust analytics, Sentry, Firebase, WebRTC, OTA updates, in-browser voice (ONNX Runtime Web) |
| DevOps | Kubernetes (StatefulSets, Services, Networking), Docker, Helm, GitLab CI/CD, GitHub Actions, Terraform, AWS (EC2, Amplify, RDS, DynamoDB, VPC), Linux, Nginx, Database Replication, Grafana, InfluxDB |
| Security | HSM, JWE, OAuth 2.0 + mTLS, JWT, Keycloak, TLS, cryptography, QWAC and QSEAL compliance |
| Soft Skills | Leadership, Abstract Thinking, Team Collaboration, Documentation, Management |
