Selected Projects

The Challenge

Crypto investors lacked tools to automatically optimize yield across DeFi protocols — manual strategies were slow, risky, and left money on the table. I set out to build an AI-driven protocol that could autonomously manage liquidity and maximize returns on Waves blockchain.

What I Did

• •Took the product from idea to production — secured funding, assembled and led a cross-functional team of 5
• •Architected end-to-end infrastructure: Ride smart contracts, AI agents (LangChain/LangGraph with RAG), Telegram bots, and web frontend
• •Drove product strategy through user research and iterative testing, prioritizing features based on real DeFi market demands
• •Grew the active user base to 60+ investors through community building and product-led growth

The Challenge

Pygma, an AI-powered Instagram growth tool, had zero data infrastructure — no analytics, no attribution, no way to measure what drove product growth or inform ML model training. I was brought in as the first data hire to build everything from scratch.

What I Did

• •Built the entire analytics stack from zero: BigQuery warehouse, Airflow orchestration, dbt transformations, Mixpanel product analytics, GTM and GA4 event tracking
• •Collected and processed 20TB of Instagram profile data via APIs to power ML model training datasets
• •Designed and shipped an attribution system matching user events to Meta/Google Ads campaigns — gave the growth team clear visibility into marketing ROI
• •Created executive dashboards that became the primary decision-making tool for product, growth, and operations teams

The Challenge

DeFi investment decisions across multiple blockchains required pulling fragmented data from dozens of sources — there was no unified view. I built a private analytics platform to aggregate cross-chain data and enable real-time investment analysis.

What I Did

• •Designed and managed data warehousing for 6 blockchain networks (EVM chains, Tron, Near) with real-time analytical data marts
• •Deployed and maintained the full infrastructure stack: Clickhouse, Airflow, Grafana, and multiple blockchain nodes
• •Built ETL pipelines ingesting data from 8+ sources for smart contract labeling and cross-chain opportunity analysis

The Challenge

A fast-growing eCommerce agency needed a data warehouse that could handle massive daily volumes across orders, inventory, and user analytics — while keeping queries fast and storage costs under control. The existing infrastructure couldn't scale.

What I Did

• •Designed scalable data warehouses on Yandex Cloud processing 50M records daily with sub-second query performance
• •Engineered automated ETL pipelines (Airflow, dbt, Nifi) and built a data quality framework with Great Expectations
• •Optimized storage architecture, cutting costs by 30% without sacrificing query speed
• •Scaled the data engineering team — ran training sessions and conducted technical interviews

The Challenge

Enterprise clients in banking and retail needed reliable, high-throughput data integrations across hundreds of database tables — with strict SLAs, complex data lineage, and zero tolerance for downtime. I joined as part of Glowbyte's consulting team and grew from junior to a key contributor on major client engagements.

What I Did

• •Built and maintained 13 enterprise data integrations managing 1,500+ database tables across banking and retail clients
• •Developed real-time analytics tools and custom dashboards, reducing data processing latency by 30%
• •Established CI/CD pipelines for seamless deployment across multiple client environments
• •Contributed to high-profile projects: SIBUR's data-driven transformation and Lemana PRO's AI-powered demand forecasting