Computer Science graduate specializing in backend engineering, cloud technologies, and applied AI. Experienced in designing and building scalable, production-ready systems using Node.js, TypeScript, PostgreSQL, Redis, RabbitMQ, Docker, and AWS. Developed AI-powered applications, including RAG pipelines and multi-agent systems, with a focus on performance, reliability, and maintainability. Strong foundation in distributed systems, cloud architecture, and DevOps practices, complemented by open-source contributions and a passion for building software that solves real-world problems.

• 🔭 Building FinBot — a financial-intelligence API. Terraform IaC on DigitalOcean, full CI/CD + DevSecOps (Gitleaks, Trivy, Datadog APM), and 139 integration tests against real infrastructure (no mocks). Under the hood it also runs a multi-agent RAG pipeline (FinBERT + Pinecone).
• 🧱 Into the hard parts of backend: event-driven architecture, the outbox pattern, PostgreSQL sharding, CQRS / event sourcing, consensus (Raft), and building services that stay correct under load.
• 🛡️ Penetration testing & bug-bounty background — I carry security into backend and infrastructure (DevSecOps, OWASP Top 10), and into LLM systems where it matters (RAG guardrails, prompt-injection defense).
• 💬 Ask me about Node.js at scale, event-driven systems, microservices, Docker, or CI/CD.
• 📫 Reach me on LinkedIn · Email

Languages

Backend & APIs

Data & Messaging

DevOps & Cloud

AI / LLM (applied) — RAG & agents layered on top of backend services

• QuickBite — Microservices food-delivery backend, split across core · order · analytics services. JWT auth + RBAC, order-lifecycle state machine, Kashier payments, WebSockets, regional PostgreSQL sharding, and a RabbitMQ outbox + worker. TypeScript · Node.js · Go · MongoDB · PostgreSQL · Redis.
• FinBot — Production financial-intelligence API. Terraform IaC on DigitalOcean, CI/CD, DevSecOps (Gitleaks · Trivy · Datadog APM), and 139 real-infra integration tests (no mocks). Multi-agent RAG pipeline (FinBERT + Pinecone) under the hood.
• redis-nodejs — Redis reimplemented from scratch in Node.js to understand its internals (caching, queues, sessions, protocol).
• MASAR — Mentorship platform on an event-driven backend. Express 5 · TypeScript · Prisma · PostgreSQL · RabbitMQ · Redis · Socket.io, with a FastAPI microservice for the AI features.
• agentic-ai-from-claude-code   — Agentic-AI patterns reverse-engineered from Claude Code's architecture (planner/tool loops, typed error handling).

Merged production code into widely-used developer-tooling, infrastructure, and AI projects (badges show live star counts):

• Floci   — Fast, open-source AWS local emulator (Quarkus / Java). 7 merged PRs on core infrastructure: Amazon MQ broker control plane (RabbitMQ-backed), ElastiCache rollback fix, MemoryDB ACL / user-auth model, and DocumentDB emulation (DocDbService, DocDbQueryHandler).
• Continue   — Open-source AI code assistant / IDE autopilot. Contribution shipped in v1.37.0.
• Unstract   — LLM-driven document extraction platform (Python / Django). PR #2158 — added a native Mistral embedding adapter (mistral-embed) via LiteLLM, so token cost tracking works out of the box.
• ApexYard   — AI-agent SDLC framework (Shell / Markdown). PR #810 — fixed the /roadmap skill's target-repo resolution and tracker exit-code handling.
• Strix   — Autonomous AI penetration-testing agents. Agent-tooling contributions — a fit for my pentest + security background.
• TradingAgents   — Multi-agent LLM financial-trading framework. Overlaps with the agentic-RAG side of FinBot.