Professional Summary

Software Engineer focused on backend architecture, distributed systems, full-stack products, and production AI. Builds reliable services and agentic systems with TypeScript/Node.js, PostgreSQL, Redis, Docker, AWS, and modern LLM tooling. Hands-on with agent planning, tool calling, RAG, memory, multi-agent workflows, structured outputs, human-in-the-loop controls, evaluation, observability, retries, and long-running execution. Founder and lead engineer of Ryuksaidso, an agent reliability and control-plane platform. Uses AI coding and research tools to accelerate implementation, debugging, documentation, refactoring, and test generation, followed by human review, automated validation, security checks, and architecture verification before changes ship.

Core Technical Skills

Languages

TypeScriptJavaScriptSQLPython

Backend

Node.jsExpress.jsFastifyNestJSREST APIsWebSocketsMicroservicesEvent-Driven Architecture

Frontend

React.jsNext.jsReduxTailwind CSS

Data

PostgreSQLMongoDBRedispgvectorPrisma

AI / Agents

AI AgentsMulti-Agent SystemsPlanningTool CallingMemoryRAGEmbeddingsStructured OutputsHITLLangChainLangGraph

LLM Engineering

Local LLMsSecure LLM HarnessesFine-TuningLoRA/QLoRAQuantizationGGUFInference OptimizationModel Routing

Cloud / DevOps

AWS EC2IAMDockerDocker ComposeNGINXCI/CDLinuxPrometheusGrafanaLoki

Engineering

System DesignDistributed SystemsAPI SecurityAuthenticationRBACRate LimitingCachingObservabilityPerformanceDSA

AI-Augmented Development

LLM-assisted codingDebuggingTest generationDocumentationImplementation accelerationVerificationHuman-in-the-loop review

Professional Experience

Ryuksaidso | Product & Engineering

Live product: ryuksaidso.online

Production-oriented agent reliability and control plane designed to operate AI agents with production software discipline. The platform centers on planning, tracing, approvals, evaluations, versioning, knowledge retrieval, and multi-provider failover.

• Control Plane: Next.js/React dashboard for users, workspaces, agents, tools, providers, runs, approvals, traces, analytics, profiles, and administration.
• API Layer: Stateless TypeScript/Node.js services with validation, authentication, RBAC, rate limiting, provider routing, health/readiness endpoints, and horizontal-scaling support.
• Agent Runtime: Planning, task decomposition, tool calls, memory/state, structured outputs, retries, human approval, execution traces, and recovery.
• Async Compute: Redis + BullMQ workers for long-running agent execution, browser automation, ingestion, scheduled jobs, retries, and workload isolation.
• Data Layer: PostgreSQL + Prisma for transactional data, pgvector for semantic retrieval, and Redis for queues, caching, locks, and coordination.
• LLM Layer: OmniRoute, Ollama, and OpenAI-compatible providers with local/cloud routing, quantized/fine-tuned model support, and controlled tool execution.
• Production Infrastructure: Dockerized services on AWS EC2 behind NGINX with IAM, security groups/firewall controls, service isolation, health checks, and controlled network exposure.
• Reliability & Safety: Human-in-the-loop approval gates, restricted tool permissions, policy enforcement, structured validation, auditability, retries, timeouts, and persistent execution state.
• Observability: Prometheus metrics, Grafana dashboards, Loki logs, audit trails, agent traces, worker visibility, and operational health monitoring.

Education