Justin Lord — Resume

About

11 years building software — started at 8. I specialize in autonomous AI agent systems, trading infrastructure, and production LLM integrations. I build and ship AI-powered products end-to-end: architecture, code, deployment, revenue. Right now I build and sell AI systems to businesses full-time through EmpusaAI. Deep experience in agent safety, real-capital autonomous execution, and scaling multi-agent deployments where failure means losing money, not just a bad demo.

Technical Skills

Languages

Python, JavaScript/TypeScript, Solidity, SQL

AI / LLM

Claude API, OpenAI API, prompt engineering, inference pipelines, multi-agent orchestration, RAG, eval frameworks

Backend / Infra

Node.js, FastAPI, Express, Supabase, PostgreSQL, MongoDB, Redis, Railway, Vercel, WebSockets

Frontend

React, Next.js, Tailwind CSS

Web3 / Crypto

Sui Move, Solana, Monad, EVM, smart contracts, on-chain agents, Polymarket CLOB API

Agent Systems

Autonomous execution loops, safety guardrails, failure recovery, workflow orchestration, monitoring at scale

Experience

EmpusaAI — empusaai.com · Mumbai, Remote

Build and sell custom AI systems to businesses end-to-end. Sole engineer across full product lifecycle.

Built autonomous trading bots on Polymarket handling real capital: entry logic, position sizing, exit execution, CTF contract redemption — zero human intervention. Safety layers for 24/7 unsupervised operation.
Deployed 1,200+ AI agents on Monad blockchain with pooled capital, committee-based trading decisions, and real-time monitoring dashboard.
Shipped empusaai.com — consumer AI product with Claude API inference pipeline, Stripe paywall, and Supabase backend. Live with paying users.
Built prediction market arbitrage analysis system — real-time cross-platform price comparison between Kalshi and Polymarket with automated opportunity detection.
Built AI lead generation backends with Supabase, PDL/Lusha API integrations, automated scoring, and client-facing dashboards for B2B clients.
Ship weekly. Multiple concurrent client projects and products at all times.

Freelance & Contract · Remote

Built AI automation systems for service businesses: missed call recovery, lead follow-up, scheduling, client management dashboards.
Developed trading systems integrating prediction market APIs, real-time data feeds, and automated execution strategies.
Created content automation tools using social media APIs, engagement velocity analysis, and trend detection pipelines.

Independent · Started at age 8

Progressed through web development, backend systems, APIs, databases, then AI and blockchain over 7 years of self-directed engineering.
Built and shipped multiple projects across web, mobile, and blockchain.

Live Projects

↗

Consumer AI

EmpusaAI

AI tool that tells people what makes them irreplaceable by AI. Claude API inference, Stripe payments, live counter, shareable result cards.

React · Claude API · Stripe · Supabase · Vercel

↗

Agent Platform

Fund Agent Dashboard

Autonomous AI venture capitalist on Monad. Multi-agent committee deliberation on every trade. 47 agents, 1,247 MON pooled, strategy builder with YAML config.

React · Monad · Multi-Agent · WebSockets

↗

Web3 · Hackathon

SealForge

AI agent marketplace on Sui. Agents create encrypted intelligence reports via Seal threshold encryption, store on Walrus, sell on-chain autonomously.

Sui Move · React · Seal Encryption · Walrus

↗

Trading · Research

Arbitrage Breakdown

Visual explainer of prediction market arbitrage — how $40M+ was extracted from cross-platform price discrepancies between Kalshi and Polymarket.

React · Data Visualization · Polymarket API

↗

Client Tool

FenceQuote Pro

Map-based quoting tool for fencing businesses. Draw fence lines on satellite imagery, get instant cost estimates with material presets.

React · Mapbox · Geospatial · Custom Pricing

↗

Game

Higbrokes

Browser-based game built and deployed on Railway. Because sometimes you just build something fun.

JavaScript · Railway

Why Agent Systems

I've spent years building autonomous agents that handle real money with no human in the loop. That forces you to think about safety, failure modes, observability, and guardrails as survival — not theory. The hardest part of agent systems isn't the AI. It's making sure the AI doesn't break things at 3am when nobody's watching. That's the problem I want to solve at scale.