Project
A collection of projects showcasing my work and expertise in various domains.
AI animation studio where users describe a scene in plain language, a language model writes it as Canvas 2D JavaScript, and the browser plays the code like a film, ready to export as MP4 with no cameras, editing, or video model.
Role
Full Stack AI Product Engineer
Technical Highlights
- Built a prompt-to-animation pipeline where users direct subject, mood, lighting, and camera in plain language and an LLM generates complete Canvas 2D JavaScript scenes
- Implemented in-browser frame rendering with one-click MP4 recording, plus downloadable standalone HTML files containing the generated animation code
- Ran generated code safely in an isolated, network-blocked preview frame, with private storage so only the creating account can open its animations
- Designed a pay-as-you-go credit model with prepaid packs, upfront pricing for lengths from 5 seconds to 5 minutes, automatic refunds for failed generations, and free credits on email verification
- Integrated cryptocurrency checkout through NOWPayments for credit purchases without a subscription
- Laid the groundwork for an upcoming Three.js mode that extends the same prompt-to-code workflow to 3D animation
Impact
Lets creators and storytellers produce video-like animations by writing like a director instead of learning a video editor or coding, turning a single sentence into a shareable MP4 through code.
- AI
- LLM
- Code Generation
- JavaScript
- Canvas 2D
- Three.js
- MP4 Export
- Sandboxed iframes
- NOWPayments
Paint the Escape is a 2D browser game where every jump brings a dark city to life. Players leap across rooftops, leaving bursts of color, glowing windows, and blooming flowers while escaping a wave of darkness.
Role
Independent Game Developer
Technical Highlights
- Designed hold-and-release jump controls with charge-based distance so longer holds travel farther across rooftops
- Implemented precise landing detection with combo scoring, including Perfect landings that create larger paint splashes
- Built a visual transformation system that wakes a charcoal city into coral, cyan, violet, and gold as players land
- Tuned increasingly challenging jumps while keeping runs short, readable, and playable on both desktop and mobile
- Added a postcard save so players can keep a snapshot of the city they painted after each run
Impact
The game combines simple hold-and-release controls with precise landings, combo scoring, and increasingly challenging jumps. Designed for desktop and mobile, it pairs fast-paced gameplay with a vibrant visual transformation that makes every run feel rewarding.
- HTML5
- JavaScript
- 2D Game Development
- Game Design
- Platformer
- Endless Runner
- Mobile
- Desktop
Sponsored discovery directory for Lovable-built websites where builders pay to claim and climb ranks, browse by category and today’s listings, and get discovered through paid placement instead of votes or waitlists.
Role
Full Stack Product Engineer
Technical Highlights
- Built an end-to-end sponsored ranking product on TanStack Start, React 19, and Tailwind with public directory surfaces for all-time ranks, category boards, today’s joiners, featured placements, and website detail pages
- Designed a balance-based ranking model in PostgreSQL where higher paid balances rise higher, with tie-break timing, category partitions, and a $2 listing floor aligned to Paystack USD minimums
- Implemented a public claim-and-pay flow: URL and category intake, Open Graph metadata enrichment, bid preview, Paystack checkout, payment verification, then secure listing credit
- Modeled the data layer with websites, ranking accounts, ranking payments, featured bookings, categories, and RLS policies for anonymous public reads with privileged writes through security-definer RPCs
- Optimized directory and profile performance with count-based rank queries, shared directory chrome, searchable category comboboxes, lazy logo loading, and reduced overfetching on category and detail routes
- Shipped discovery UX for mobile and desktop: claim UI, podium list cards, category filters with search, sidebar today’s rank, and SEO metadata for home, category, today, and detail pages
Impact
Gave Lovable builders a paid discovery channel to get listed, rise by investment, and be found by category — with real checkout, secure server-side payments, and a production-ready ranking board ready for community marketing.
- TanStack Start
- TanStack Router
- React 19
- TypeScript
- Vite
- Tailwind CSS
- Supabase
- PostgreSQL
- Row Level Security
- Supabase Edge Functions
- Paystack
- Payments
- Server Functions
- Zod
- SEO
- Lovable
AI-powered voice sales platform that conducts natural inbound and outbound calls, follows approved sales scripts, answers customer questions, handles objections, qualifies leads, and coordinates follow-ups or live transfers to human representatives.
Role
Full Stack AI Engineer
Technical Highlights
- Architected an end-to-end voice pipeline connecting Twilio telephony, speech-to-text, LLM reasoning, and ElevenLabs text-to-speech for natural customer conversations
- Designed AI sales conversation logic that follows approved scripts, handles common objections and rebuttals, answers product questions, and determines the next best action
- Implemented RAG-based retrieval across product information, FAQs, policies, sales materials, and authorized live business data such as pricing and availability
- Built inbound and outbound calling workflows with lead qualification, call outcome capture, appointment scheduling, and automated SMS and email follow-ups
- Integrated CRM and lead management workflows to store conversation summaries, customer interest, objections, and actionable next steps
- Implemented warm call transfers for customer requests, high-value leads, and conversations requiring human sales expertise
- Designed the voice agent architecture for configurable AI models based on response quality, latency, reliability, privacy, call volume, and operating cost
Impact
Enabled businesses to automate scalable sales conversations while preserving accurate company messaging, structured lead capture, timely follow-ups, and seamless human escalation for qualified or complex opportunities.
- Next.js
- TypeScript
- AI
- LLM
- Voice AI
- Twilio
- ElevenLabs
- Speech-to-Text
- Text-to-Speech
- RAG
- CRM Integration
- Real-Time Communication
- Vercel
AI-powered music creation platform that helps creators generate original music, lyrics, melodies, and complete productions through intelligent AI workflows while providing tools for editing, organizing, and publishing creative projects.
Role
Full Stack AI Engineer
Technical Highlights
- Architected a scalable asynchronous music generation pipeline with queue-based processing, job lifecycle management, retries, progress tracking, and failure recovery for long-running AI generation tasks
- Built the platform using Next.js, TypeScript, and Supabase, leveraging PostgreSQL, Authentication, Storage, Edge Functions, and Row-Level Security for a secure multi-user SaaS architecture
- Designed a production-ready backend architecture supporting concurrent music generation requests, real-time status updates, optimized database operations, and efficient resource utilization
- Integrated AI-powered music generation workflows capable of producing complete songs, instrumentals, vocals, lyrics, and multiple musical styles through configurable generation pipelines
- Developed a comprehensive project management system with generation history, versioning, favorites, search, filtering, and cloud-based asset management
- Implemented subscription, credit-based billing, secure authentication, and user quota management to support commercial-scale AI music generation services
- Optimized frontend performance with Next.js App Router, server components, streaming, caching, and real-time Supabase subscriptions to deliver a responsive user experience
Impact
Empowered musicians, creators, content producers, and businesses to transform ideas into professional-quality music within minutes, dramatically accelerating creative workflows while making high-quality music production more accessible.
- Next.js
- TypeScript
- Supabase
- PostgreSQL
- AI
- LLM
- Music AI
- Audio Processing
- Stripe
- Crypto payment
- Vercel
- Tailwind
- shadcn/ui
AI scheduling platform for marketing content that generates, organizes, schedules, and automatically publishes social media content using AI-powered workflows and connected social accounts.
Role
Full Stack AI Engineer
Technical Highlights
- Built an AI-powered marketing automation platform that transforms a single content idea into weeks or months of scheduled social media content
- Integrated Groq Llama 3.3 70B and AI SDK to generate high-quality marketing posts with configurable tone, audience targeting, and content strategies
- Implemented secure Bluesky account connection and OAuth-based authentication for automated publishing workflows
- Designed a scheduling engine that automatically generates publishing calendars and manages future content across customizable posting frequencies
- Created a content management dashboard for reviewing, editing, scheduling, and monitoring AI-generated marketing campaigns
- Developed automated publishing infrastructure with retry mechanisms, publishing logs, token management, and delivery tracking
- Implemented timezone-aware scheduling and content batching to support global marketing campaigns and consistent audience engagement
Impact
Enabled marketers, creators, and businesses to automate content planning, generation, scheduling, and publishing from a single platform, significantly reducing manual effort while improving content consistency and audience reach.
- Next.js
- TypeScript
- Supabase
- PostgreSQL
- Groq
- AI SDK
- Bluesky API
- OAuth
- Vercel
- Cron Jobs
- Tailwind
- shadcn/ui
Advanced cryptocurrency trading platform offering spot trading, instant swaps, copy trading, and multi-router DeFi liquidity aggregation.
Role
Senior Full Stack Engineer
Technical Highlights
- Developed core trading platform functionality including copy trading, spot trading, and instant trading workflows
- Designed and built a Meta DEX Aggregator Engine utilizing multiple routing strategies to optimize swap execution
- Implemented high-performance routing logic to deliver faster transactions, improved liquidity access, and reduced trading costs
- Built backend services and APIs supporting real-time trading operations and portfolio synchronization
Impact
Improved trading efficiency and execution quality through intelligent routing, enabling users to access optimal liquidity across multiple decentralized exchanges.
- Next.js
- TypeScript
- Node.js
- PostgreSQL
- Web3
- DeFi
- DEX Aggregation
- Copy Trading
- Smart Contracts
- Trading Systems
Prediction market platform enabling users to participate in event-based markets and leverage social trading strategies powered by blockchain technology.
Role
Senior Full Stack Engineer
Technical Highlights
- Contributed to full-stack development of the prediction market platform across frontend, backend, and data services
- Integrated third-party prediction market providers and external market data sources, including Polymarket-compatible endpoints
- Implemented one-click copy trading functionality allowing users to automatically replicate successful trading strategies
- Built real-time market synchronization, trade execution, and portfolio tracking systems
Impact
Enhanced user engagement and trading accessibility by simplifying prediction market participation and enabling automated strategy replication.
- Next.js
- TypeScript
- Node.js
- PostgreSQL
- Web3
- Prediction Markets
- Copy Trading
- API Integrations
- Real-Time Data
- Blockchain
Web2.5 blockchain infrastructure ecosystem designed to bring tokenization, digital asset management, transaction tracking, and Web3-style functionality into familiar Web2 application environments.
Role
Lead Full Stack Engineer
Technical Highlights
- Engineered core full-stack infrastructure for blockchain-enabled applications, integrating token, transaction, account, and digital asset workflows across the DFS Chain ecosystem
- Developed blockchain explorer and transaction-tracking functionality for monitoring asset activity, transaction records, hashes, and ecosystem operations
- Built wallet and account-adjacent infrastructure enabling users and applications to manage digital assets through a simplified Web2-oriented experience
- Designed token lifecycle and ecosystem service architecture supporting token creation, transfers, application-level utility, and interconnected digital asset workflows
- Developed and integrated APIs connecting frontend applications with blockchain and ecosystem services, enabling consistent data access and transaction processing across multiple products
- Contributed to ecosystem applications spanning token launches, exchanges, asset management, transaction visualization, and interconnected Web2/Web3 experiences
- Designed reusable full-stack components and service integrations to support the expansion of multiple applications operating within the DFS Chain ecosystem
Impact
Contributed to the development of a scalable Web2.5 infrastructure ecosystem that simplified blockchain-style functionality for applications and users while enabling tokenization, transaction tracking, digital asset management, and cross-application connectivity.
- Supabase
- PostgreSQL
- Next.js
- TypeScript
- Blockchain
- Web3
- DeFi
- Tailwind
- DApp Architecture
- Token Economy
RAG-based AI consultant that answers user questions about the DIFINES ecosystem using markdown knowledge resources, Supabase pgvector search, and Groq-powered LLM responses.
Role
Full Stack AI Engineer
Technical Highlights
- Built a RAG chatbot for the DIFINES consultant page using markdown documents as the knowledge source
- Implemented Supabase pgvector storage and similarity search for retrieving relevant ecosystem documentation
- Integrated Groq Llama 3.3 70B for fast AI responses and Gemini embedding-001 for document embeddings
- Created an ingestion workflow to chunk markdown files, generate embeddings, and store searchable knowledge in PostgreSQL
- Designed the assistant UI to match the existing DIFINES landing page style and provide a smooth chat experience
Impact
Improved user understanding of the DIFINES ecosystem by enabling instant AI-powered answers based on verified project documentation.
- React
- TypeScript
- Supabase
- PostgreSQL
- pgvector
- AI SDK
- Groq
- Llama 3.3 70B
- Gemini embedding-001
- RAG
- Markdown Knowledge Base
- Tailwind









