Featured Projects
Real products and AI systems built to scale, operate reliably, and actually solve business problems.
About Me
Product Engineer & AI Architect. I design, build, and ship entire systems — from architecture and AI to frontend and operations.
- â—†Built and scaled the Loveepil platform across web, admin, analytics, and AI services.
- â—†Delivered 20+ projects (14 from scratch), handling everything from architecture to production release.
- â—†Led end-to-end product engineering, including localization, payments, and system reliability.
How I Work
- →Architecture first: I define clear boundaries, data flows, and contracts before writing code.
- →Product-minded: I focus on connecting business goals with the technical reality of the platform.
- →Full ownership: I build, ship, and operate systems, making sure they actually work in production.
Engineering Principles
Career Timeline
I work in a hybrid model: long-term core roles combined with architecture and product engagements.
Stack
Responsibilities
Owned performance and architecture of drilling analytics dashboards used by enterprise oil & gas clients. Collaborated closely with QA, designers, and PM. Conducted code reviews and designed a 7-phase refactoring playbook for a complex charting codebase.
Achievements
Cut page load from 5s to 1.5s (3x faster) by eliminating render cascades, fixing chart caching, and lazy-loading heavy modules. Restructured the codebase into a domain-driven architecture — zero regressions, no feature freeze.
Stack
Responsibilities
Sole developer of an 8-service production ecosystem across 5 countries. Three AI systems live in prod: a 21-tool AI consultant handling bookings, FAQs, and service guidance across 4 channels, an autonomous content engine generating ~70 articles/week, and embedding-powered semantic search. Plus analytics, admin panel, payments, and transactional emails.
Achievements
Deployed a production AI agent running at $0.015 per dialog with near-zero hallucinations. Achieved a 96% Mobile Performance Score. Built cross-domain analytics with phone-based attribution.
Projects
A custom AI agent using fine-tuned GPT-4.1 Mini to orchestrate 21 tools. It handles bookings, answers FAQs, and seamlessly transfers to human operators when needed. Fully observable with Langfuse.
Custom platform that tracks user journeys across 5 country domains. Built a phone-based attribution system to tie CRM bookings back to Google/Meta ads.
Stack
Responsibilities
Led frontend modernization across 3 internal products, working with QA, backend engineers, and PM. Moved the stack from legacy setups to a unified Next.js architecture. Conducted PR reviews and improved security by integrating automated vulnerability checks into CI/CD pipelines.
Achievements
Rewrote the authentication layer across 3 products, eliminating critical security vulnerabilities in session handling. Migrated 50K+ lines to TypeScript and Next.js — reduced onboarding time for new engineers and cut regression bugs through strict type safety.
Stack
Responsibilities
Built 2 greenfield React applications from scratch for US and UK enterprise clients. Owned full frontend delivery — from architecture and component design to testing and production release.
Achievements
Delivered Modo Energy analytics platform with 40+ interactive chart widgets for the UK energy storage market. Built Zymergen data visualization tool handling large biotech datasets with complex filtering and drill-down interactions.
Skills
Technologies and tools I use to build scalable, high-performance applications.
