I design and engineer LLM-powered products, scalable APIs, and interfaces that feel effortless. Turning complex AI into products people actually use.

I'm a software engineer and DevOps architect with 12+ years of experience building and scaling products across Python, Java, JavaScript, and PHP, using frameworks like Spring, Django, FastAPI, Angular, React and more. I've designed and operated cloud-native platforms on AWS and GKE with Kubernetes and Docker, with a strong focus on reliability, performance, security and delivery speed. My current focus is AI-first product engineering: building machine learning and agentic systems with PyTorch, NumPy, Pandas, and LangChain to turn cutting-edge models into practical business outcomes.
Architected and operated cloud-native platforms across AWS and GKE using Kubernetes and Docker, while designing and developing robust Java applications and production-ready AI agents with Python, PyTorch, and LangChain.
Engineered ship chartering and freight execution platforms on Kubernetes across cloud and on-prem clusters, building high-throughput services in Python, Java, and PHP and delivering Angular/React interfaces for booking, scheduling, and end-to-end cargo visibility.
Owned the full lifecycle of on-demand B2B services, from product management and requirements discovery through architecture, development, testing, maintenance, deployment, and production monitoring to ensure reliable, scalable delivery.
Architected and developed custom WordPress plugins and themes, delivering scalable, maintainable solutions tailored to client workflows, performance goals, and brand requirements.
LLM pipelines, RAG systems, and intelligent features embedded directly into your product. From proof-of-concept to production-grade infrastructure.
An AI support agent that works a shared inbox end to end. Hazel triages incoming conversations, drafts replies grounded in connected systems — Shopify orders, Stripe payments, returns, and a help center — surfaces the exact records behind every answer, and flags the cases that genuinely need a human. Routine tickets get resolved automatically while the team keeps full oversight.
An enterprise knowledge assistant that answers any question about your company by searching across every connected source — Slack, Google Drive, Notion, Confluence, Jira, GitHub, and an SOP library — and citing the exact documents behind each answer. It handles onboarding questions, policy lookups, decision history, and meeting summaries, with a live feed of what the rest of the team is asking.
A control plane that keeps a fleet of AI agents safe and accountable. Warden enforces live guardrails on every action, redacts PII, gates spend, and routes high-risk steps to human approval — while a flight-recorder audit log captures every decision, verdict, and record. It includes a policy studio for authoring rules and a red-team & evals suite for stress-testing agents before they ship.
An intelligent document desk for the back office. Manila reads everything that lands in the mailroom — invoices, purchase orders, contracts and NDAs, insurance claims, HR forms — classifies and routes each one with a confidence score, auto-files the clear cases, and escalates the rest for human review. It ships with a rulebook, configurable document types, an approvals queue, and throughput analytics.
A unified platform for managing, training, and interacting with AI bots. Users can spin up bots with custom personas, equip them with selectable tool sets (web search, code execution, file access, APIs), and chat with them in real time. Includes a bot registry, per-bot memory and context management, tool permission controls, conversation history, and an analytics dashboard tracking usage and performance across the entire bot fleet.