Harrison Wachira@harryiieng63
5
Events attended
2
Submissions made
Kenya
5+ years of experience
About me
Electrical and Electronics Engineering professional with a strong interest in AI and data analytics. I enjoy building data-driven projects, specifically in the sports and football betting markets. Currently exploring the intersection of electronics and AI to create innovative solutions. Excited to collaborate, learn, and build during hackathons.
🤓 Latest Submissions
Strike Sentry
Strike Sentry is an AI-powered options trading agent built for the Alpaca AI Trading Agents Hackathon. It implements a cash-secured put strategy on Alpaca's Paper Trading API, using Groq for reasoning (with a deterministic keyword fallback for reliability) and Alpaca's official CLI for order submission — satisfying the hackathon's requirement to use Alpaca's Trading API and CLI rather than raw REST calls for execution. The agent's pipeline works in three stages. First, it pulls live account positions directly from Alpaca. Second, given a target ticker and strike, it queries Alpaca's real options chain snapshot feed, filters for contracts with genuine tradable quotes (not just theoretically valid strikes), and returns a full trade proposal including strike, expiry, premium, cash required, net cash at risk, and breakeven. Third, it submits the proposed order through Alpaca's official CLI binary, authenticated via `alpaca profile login`, rather than a raw REST call. Every order is checked against real-time buying power before submission, and the backend enforces a paper-only execution guardrail by default. The system was built and tested end-to-end on Node.js/Express, deployed live on Render, with a Groq-powered decision layer that defaults to "hold" under uncertainty rather than guessing. During testing, the full pipeline was verified against Alpaca's live paper trading servers: real options chain lookups, accurate strategy math, successful CLI authentication, and order submission reaching Alpaca's production paper API end-to-end.
4 Sep 2026
LogLens — AI-Native Log Triage Dashboard
LogLens is an AI-native log triage dashboard built for DevOps engineers and SREs who need to quickly make sense of large volumes of server logs during an incident. Instead of manually scanning through pages of raw log output, users paste their logs into the app and click "Run Triage." An AI model analyzes each log entry and classifies it into one of three severity categories: Critical, Warning, or Info. Each classified entry is displayed as a color-coded card — red for critical failures like database timeouts or payment gateway errors, yellow for warnings such as high memory usage, and green for routine informational events like successful logins. This visual hierarchy lets engineers immediately spot the issues that need urgent attention without reading every line. The project was built using native.builder's agent-driven workflow, going from a plain-English description of the desired app directly to a deployed, functional product. This demonstrates how AI-native development platforms can compress the traditional build cycle for operational tooling — turning an idea into a working incident-response aid in minutes rather than days.
10 Aug 2026
