expertise
What I can do, and where I have actually done it.
Ordered by where I spend my time now: AI and agentic systems first, then the backend, frontend, infrastructure, security and Web3 work that surrounds them. Technologies are evidence for capabilities, not a catalogue.
Levels describe how I have used a technology, not a score. Production: shipped and operated in a real product. Hands-on: built with it in projects or prototypes. Exploring: studying or experimenting.
proof, not logos
- Shipped an AI concierge → CRM → voice-agent → telephony pipeline with tool-calling, orchestration and webhooks, verified with a live call. Tellco AI
- Architected the LLM-integration layer behind an adaptive-learning platform with streaming and voice. SchoolExl / TrainExl
- Built the text-to-speech, voice catalogue and translation capabilities of a workspace-scoped voice API. Voice Labs
- Built a human-in-the-loop triage assistant with guardrails in code and an adversarial eval set (19/26 → 24/26). Lab: inbox triage
LLM engineering
Production- LLM API integration (OpenAI, Anthropic)
- Prompt engineering
- Structured outputs
- Streaming LLM responses
- Context management
- Conversation state management
- Model-to-UI communication
- LLM-powered applications
Agents & agentic workflows
Production- AI agent development
- Tool and function calling
- Multi-step orchestration
- Agent-to-tool communication
- Workflow automation
- Webhooks and external tool integrations
- Agent state managementhands-on
- Human-in-the-loop workflowshands-on
- MCPhands-on
Voice & conversational AI
Production- Voice AI
- Conversational AI
- Real-time AI interactions
- WebRTC
- FreeSWITCH
- Voice agent integrations
- Multilingual AI experiences
- Speech-to-text
- Text-to-speech
- Streaming audio workflows
- LiveKit
- Pipecat
Retrieval & knowledge systems
Production- Retrieval-augmented generation (RAG)
- Document processing
- Knowledge base integration
- Context retrieval
- Grounded responses
- Contextual conversation handling
- Semantic searchexploring
Evaluation & reliability
Hands-on- Agent evaluation
- AI workflow testing
- Response-quality evaluation
- Tool-calling validation
- Agent benchmarking (τ-bench)
- AI system observability
- Error handling and recoveryproduction
- Reliability of LLM-powered workflowsproduction
ML exposure
Hands-on- PyTorch
- Whisperproduction
- Python-based AI pipelines
- ML inference integrationproduction
- AI/ML service integrationproduction
AI-native engineering
Production- Claude Code
- Cursor
- AI-assisted development workflows
- AI-assisted code review
- AI engineering productivity
proof, not logos
- Carried a 250+ endpoint platform over a 47-table Postgres schema and 6-tier RBAC through a zero-downtime cutover. Tellco AI
- Designed an 11-table event-sourced Postgres schema with order and address snapshots and OTP-verified delivery. Lab: PicknPay
- Prototyped a stateless realtime backplane: Socket.IO over Redis pub/sub across API replicas. Lab: video backplane
Languages
Production- TypeScript
- JavaScript
- Python
- SQL
- Gohands-on
- PHPhands-on
Frameworks & runtimes
Production- Node.js
- Express.js
- Fastify
- NestJS
- Python (FastAPI, Django)hands-on
- Bunhands-on
- Honohands-on
- Laravelhands-on
API & application engineering
Production- REST API design
- API architecture
- Authentication and authorization
- JWT
- RBAC
- Request validation
- Webhooks
- Third-party integrations
- Background processing
- Queue-driven workflows (BullMQ, RabbitMQ)
- Caching
- WebSockets and Socket.IO
- Server-side security
- GraphQLhands-on
Databases & ORMs
Production- PostgreSQL
- MongoDB
- Redis
- Prisma
- Drizzle ORM
- MySQLhands-on
Database concepts
Production- Relational and schema design
- Normalization
- Indexing
- Query optimization
- Transactions
- Event sourcing
- Pub/sub
- Data pipelineshands-on
Architecture
Hands-on- Modular architecture
- Event-driven architecture
- Scalable API design
- Asynchronous processing
- System design
- Reliability engineering
- Multi-tenant architecture
- Microservices concepts
- Distributed systems
proof, not logos
- Improved Core Web Vitals by more than 40% for 2,500 daily active users. CoreHealth
- Built a design system now reused across four products. SchoolExl / TrainExl
Languages
Production- JavaScript (ES6+)
- TypeScript
- HTML5
- CSS3
Frameworks & libraries
Production- React
- Next.js
- TanStack Start
- Angular
- Tailwind CSS
- Radix UI
- shadcn/ui
State & data
Production- Zustand
- Redux
- TanStack Query
Engineering
Production- Component architecture
- Design systems
- UI architecture
- Responsive interfaces
- Server-side rendering
- API integration
- Performance optimization
- Core Web Vitals
- Frontend testing
- Accessibility
proof, not logos
- Automated a snapshot, validate, reload and rollback-on-failure cutover with no customer-facing downtime. Tellco AI
- Built a multi-stage, multi-architecture (amd64 and arm64) Docker pipeline publishing to Docker Hub. Lab: Acquisitions API
Containers & deployment
Production- Docker
- Multi-stage builds
- Multi-architecture builds
- Nginx
- PM2
- Linux
- Production deployment
- Zero-downtime deployment
- Rollback strategies
CI/CD & automation
Production- Git and GitHub
- GitHub Actions
- CI/CD pipelines
- Automated testing
- Build and deployment automation
- Release management
- Infrastructure automationhands-on
Cloud & hosting
Production- DigitalOcean
- AWS
- Google Cloud
- Cloudflare
Monitoring & observability
Production- Pino
- Grafana
- Datadog
- Application logging
- Error monitoring
- Performance monitoring
- Production diagnostics
Secrets & infrastructure security
Hands-on- HashiCorp Vaultproduction
- Infisical
- Environment configuration
- Secrets management
- TLSproduction
- Secure deployment practices
proof, not logos
- Audited 294 endpoint and role combinations; fixed 2 critical and 5 high-severity issues (JWT CVEs, RBAC authorization leaks, exposed ports). Tellco AI
- Added rate-limiting and bot detection with Arcjet, and Better-Auth/JWT authentication, to a solo API. Lab: Acquisitions API
Application security
Production- Authentication and authorization
- Role-based access control (RBAC)
- JWT
- API security
- Input validation
- OWASP practices
- Security auditing
- Endpoint and role auditing
- Secrets management
Tooling
Hands-on- Arcjet
proof, not logos
- Shipped account-abstraction and paymaster onboarding and a Solana Farcaster mini-app. Lendbit
- Wrote a Merkle-proof ERC-1155 claim contract and a long-form review of Aavegotchi's staking diamond. Lab: Merkle claim
- Cut API latency by around 40% with subgraph queries. Web3Bridge
Smart contract engineering
Production- Solidity
- Smart contracts
- EVM ecosystem
- ethers.js
- viem
Blockchain infrastructure
Production- Subgraphs
- Blockchain indexing
- On-chain data integration
- Multi-chain development
- Solanahands-on
- SPL token flows
Web3 product engineering
Production- Account abstraction
- Smart accounts
- Paymaster integrations
- DeFi applications
- Cross-chain lending
- Blockchain data dashboards
proof, not logos
- Technical lead for an intercontinental team; set frontend conventions, CI/CD workflow and API standards, and mentored junior engineers. Experience
- Mentored 50+ developers across three programmes. Community
Foundations
Production- Object-oriented programming
- Data structures
- Algorithms
- Asynchronous programming
- Software architecture
- Design patterns
- API design
- Debugging
- Performance optimization
- Maintainable code
Practice
Production- Code reviews
- Automated testing
- Technical leadership
- Engineering mentorship