Alex Lazar
I'm a product-focused engineer with almost a decade of experience building web products for early-stage startups. I run LeetSoftware, my development shop, and I've founded or co-founded 8+ products of my own. My current work covers LLM-backed products and evaluation systems, but I have a background in financial/blockchain data pipelines.
Key achievements
- Built core product workflows for IconicShift, an AI platform for founder and executive accountability that helps founders define company strategy, plan execution and assign responsibility to team members, using company documents as context.
- Built Job Fiend, a job-search agent that automates job discovery and screening, helping job seekers find relevant opportunities and manage applications and interviews in one place.
- Resolved launch-blocking transaction failures for XGame, a blockchain betting game that handled $300k in first-week betting volume with no stuck rounds or lost balances.
- Rebuilt Paragon's trading prototype to deliver live signals and execute trades without hourly browser freezes. Processed 3,000-4,000 market-data events per second across 400-500 trading pairs.
Favorite Tech Stack
I've worked with a broader range of technologies, but these are my current favorites.
- Programming languages: TypeScript, Python, Rust.
- Backend: Django, Express.js.
- Frontend: React, anything from Tanstack, Inertia.js, React Router.
- Databases: PostgreSQL, MongoDB, Redis.
- Infrastructure & monitoring: AWS, Railway, Terraform, Docker, BetterStack, Sentry.
- AI/LLM: Langfuse, OpenRouter, Jev.
Experience
I contract for startups through my development shop, working as an architect and individual contributor. I've also led teams of up to five, managed roadmaps and made hiring decisions.
- Built core features for IconicShift, an AI platform for founder and executive accountability. Used LLMs to review uploaded documents against company strategy and turn founder-approved business goals into milestones with accountable owners.
- Rebuilt Paragon's Binance trading prototype to keep live signals and trade execution running without hourly browser freezes. Moved market-data processing to server-side services handling 3,000-4,000 events per second across 400-500 trading pairs.
- Built XGame's leaderboard and affiliate analytics from live and historical blockchain activity. Paired with the partner CTO to stabilize betting and payouts, adding transaction retries and recovery flows to resolve stuck rounds and refund players safely.
Dextrac provides infrastructure for Chainlink, a market-leading data network for blockchain applications, supplying real-world information such as market prices.
- Led development of PRISM, helping infrastructure providers track downtime and data-feed performance across 15 blockchains. Built data pipelines processing 10M+ transactions per month.
- Cut analytics page loads from minutes to 1-2 seconds with precomputed reports that stayed within five minutes of live data.
- Enabled paid access to live monitoring data by integrating payments and role-based access controls.
- Presented PRISM at Chainlink's SmartCon in Barcelona.
Toucan built infrastructure for trading and tracking carbon credits on blockchain networks. Its platform surpassed $2B in trading volume.
- Built and maintained an open-source TypeScript SDK for integrating carbon credits into applications. Contributed to tools that combine purchasing and retiring credits into one transaction, and managed community contributions.
- Built the bridge interface for transferring carbon assets between blockchain networks, and automated deployment of hundreds of carbon-credit tokens.
- Presented Toucan's developer tools at ETHAmsterdam, LFGrow and ETHDenver, and to teams at Porsche and Volkswagen.
- Freelanced for various digital content creators building automations, dashboards, A/B tests and an edtech platform that reached $1M/yr revenue with 10k+ MAU.
Side projects
Job Fiend (formerly Job Finder)
- Built an AI-powered job-search platform that discovers and screens roles against each customer's approved criteria and brings applications and interviews into one place. Added evals before releasing AI changes and per-customer spending controls.
- Publish product demos and technical explainers on AI and software engineering, and host interviews with founders and engineers about building products and running software businesses.
Education
- Mechanical Engineering Degree, University of Craiova, Romania (2018 - 2020). The most impactful class I had was for MATLAB. It convinced me to pursue a career in software.