Glib – Next.js, AWS, React

An agentic AI platform for enterprise partnership teams built on a proprietary partnership graph covering 10,000+ companies.

• Led engineering on an AI-powered partnership-discovery platform for enterprise partnership teams.
• Built the partner discovery and qualification surface that ranked candidate partners by impact and effort with explainable scoring.
• Designed and built an Anthropic Claude tool-use agent streamed to the browser that queried the internal company graph, ran live web research, and assembled qualification briefs.
• Owned the full stack across a Next.js / React / Typescript frontend, a Nest.js API on PostgreSQL with TypeORM, and two Python FastAPI services for enrichment and scoring.
• Deployed services with Docker on AWS ECS and used Playwright end-to-end tests as merge gates.
• Worked directly with product and design to turn ambiguous workflows into scoped interface states and reviewed pull requests across the team.

A US public-records and background-screening platform with automated data-acquisition agents.

• Built and maintained automated data-acquisition agents for a public-records and background-screening platform across 28 jurisdictions in 12 states.
• Implemented one Playwright browser-automation scraper and one Python parser per source against per-source mapping specifications.
• Introduced LLM-based extraction for inconsistent HTML sources with JSON-Schema-constrained model extraction, rule-based fallbacks, and validation to suppress hallucinated legal citations.
• Rebuilt a statewide agent after a portal moved to a React single-page application, including reCAPTCHA v2 automation, DOM polling, readiness checks, and a Chrome DevTools Protocol screenshot path.
• Diagnosed and fixed silent data-loss defects across the agent fleet, including false no-results, pagination truncation, and empty lazy-loaded tabs.
• Authored scraper, parser, QA, and compliance standards and peer-reviewed teammates’ agents against those checklists.

A B2B FMCG supply chain SaaS that automated supplier product-data extraction and retailer onboarding.

• Built the AI extraction pipeline behind the platform’s retailer onboarding workflow and shipped it during the first commercial deployments.
• Developed a multi-phase Azure OpenAI extraction pipeline in a Python worker service that converted supplier files into structured product data.
• Added a rule-based enrichment layer and an accuracy-evaluation harness measuring field-level extraction quality against ground truth.
• Delivered the Next.js / React frontend with real-time WebSocket progress and a Node.js / Nest.js backend on PostgreSQL with TypeORM, Redis-backed job queues, and Stripe billing.
• Packaged and deployed services to Azure Container Apps with Docker and Terraform.
• Agreed on extraction-accuracy targets and failure-mode handling with product and iterated with design on partial and low-confidence extraction UX.
• Reviewed teammates’ pull requests across all three services.

Contract engagement developing the company's connected fitness application, combining React Native, Unity, and real-time hardware inputs.

• Embedded Unity modules into the mobile application and connected them to Bluetooth and gyroscope sensor data.
• Implemented dynamic difficulty adjustment based on measured user performance.
• Developed Node.js and Firebase backend functionality for authentication, analytics, and leaderboards.

An AI video-generation SaaS that turned a written brief into a ready-to-post video in a single browser session.

• Built an AI video-generation SaaS end-to-end across editor, backend, AI integrations, and billing.
• Developed an AI agent that edited a user’s project by operating the application’s own interface.
• Engineered an in-browser video editor in React with Redux Toolkit and MobX using a Fabric.js and Pixi.js canvas with ffmpeg rendering.
• Built a Node.js / Express backend on MongoDB with Socket.io for real-time state.
• Orchestrated multiple external model providers behind one interface.
• Shipped the product to revenue with subscription billing across three paid tiers and authentication.
• Iterated with design on editor interactions and canvas states and participated in code review across frontend and backend work.

A music-industry booking and networking platform across web and mobile for musicians, venues, and promoters.

• Built and shipped a music-industry booking and networking platform across web and mobile.
• Built location- and genre-based search and matching features.
• Integrated Spotify and SoundCloud media embeds and an invite-based onboarding flow.
• Replaced fragmented email-and-spreadsheet booking workflows with a unified platform.
• Built features from written product specifications and aligned discovery and booking flows with design before implementation.
• Reviewed teammates’ pull requests as well as having my own code reviewed.

Cross-platform music fan-engagement application.

• Took over a legacy React Native codebase and led its architectural rebuild, interface redesign, and backend integration.
• Built core engagement features, including song brackets, fan profiles, and discovery challenges, on a Node.js backend.
• Shipped the rebuilt application to the App Store and Google Play.
• Led engineering delivery for the client product while combining technical direction with hands-on full-stack development.

A digital engineering company providing design, data, cloud, AI/ML, robotics, IoT, and XR solutions.

• Improved retail detection accuracy by 36% and reduced manual QA effort by 65% through an on-device computer-vision system using TensorFlow, TensorRT, custom CUDA kernels, and NVIDIA Jetson Nano.
• Delivered detection, segmentation, and 3D pose-estimation pipelines as Dockerized Flask/FastAPI services on AWS and GCP.
• Built Unity/C# prototypes for movement tracking and synthetic training-data generation, alongside Android applications using voice interfaces.
• Developed machine-learning, computer-vision, and interactive software systems for enterprise clients from research and experimentation through production deployment.