$launch --offer "summer-sale"
✓ generated 6 creatives
✓ published to 3 sources
▶ optimizing budget_
Software that does your repetitive work on its own.
Apps that turn messy data into clear decisions.
Sites that turn visitors into paying customers.
Selected projects shipped end-to-end for independent clients, from first structure and UI to backend logic, forms and integrations.
High-converting site for a personal fitness brand.
Landing + signup funnel for a dance course.
Service site that builds trust and books clients.
Fast signup page built to convert.
Gated content + Stripe payments platform.
AI-powered ads-optimization platform for media buyers, affiliates and agencies serving clients including Pepsi and Starbucks. Co-built two products from scratch in a team of 8 engineers + QA.
AI-powered ads-optimization platform for media buyers, affiliates and agencies. Co-built two products from scratch — trackboost.com (AI co-pilot for advertisers) and digital.trackboost.com (analytics & team-management dashboard) — in a team of 8 engineers + QA, serving clients including Pepsi and Starbucks.
Designed and shipped websites, booking systems and a Telegram bot for independent clients across fitness, wellness and education — owning the full delivery cycle from requirements and design to deployment and handover.
Thesis: "Integrated Development Environments for Work with Large Language Models." Designed and implemented a Retrieval-Augmented Generation (RAG) pipeline as the core technical contribution, grounding LLM responses in a custom knowledge base.
End-to-end campaign launcher for affiliate marketers: takes an offer + reference, generates images and headlines via OpenAI, assembles creatives, and pushes them live to traffic sources. Replaced hours of manual creative production per buyer.
Campaign optimization engine that analyzes live performance against affiliate-marketing rules and recommends concrete actions — budget adjustments, pausing underperforming creatives, and generating new variations.
n8n + Claude Opus pipeline auto-generating RSOC articles for Tonic at scale. Ran in production for several months, producing multiple campaigns per day with zero manual writing.
Live dashboard with charts aggregating a 70 GB MongoDB store and multiple SQL databases across all integrated traffic sources and campaigns. Cut key dashboard load times from ~30s to near-instant.
In-app AI creative editor to generate and refine ad images and copy, removing dependency on external design tools — plus an influencer-management system for agencies with talent CRM, campaign tracking and reporting.
Retrieval-augmented generation over a custom knowledge base — vector-search retrieval combined with LLM generation, built into an IDE for working with LLMs. Core technical contribution of the thesis.
A focused stack for building full-stack products, AI automations, dashboards, payment flows and integrations.
Full-Stack Developer with 3 years of production experience building AI-powered SaaS platforms end-to-end — from greenfield architecture to scaled features used by 200+ media buyers and agencies serving brands such as Pepsi and Starbucks.
Backend-leaning generalist across Node.js, NestJS, Express and Python, with strong React/Redux on the frontend and confident handling of MongoDB, PostgreSQL, Redis, BullMQ and AWS/GCP. Shipped multiple LLM-driven features in production — an autonomous campaign-launching agent and a Claude-powered content pipeline. Comfortable owning a feature from spec through deployment.
Drop me a line about the product, automation or site you have in mind — I reply faster than my CI (usually within 24h).