Software Engineer · Reality Interactive
May 2023 to Present · Middletown, CT
- Current focus: design systems and platform modernization, device telemetry, AWS infrastructure, and applied AI.
Nick D'Agosto
Software engineer building AI products, real-time device platforms, and production AWS infrastructure.
selected work
-
Led full stack, AI enabled development and launch of a private mortgage recruiting marketplace. Built the matching engine with eligibility filters, configurable scoring, confidence weighting, and run diagnostics, using Claude on Amazon Bedrock to explain matches and interpret lenders' geographic preferences, with rule based fallbacks. Also built a 14 tool, read only MCP server with Clerk OAuth, scoped credentials, rate limiting, and per-call audit logs.
Launched to production, recording 531 matches and 99 conversations through October 2026.
-
bitShuttle portal redesign
Led the portal redesign, building a 36 component Vue 3/TypeScript design system with five themes, then rebuilding navigation, tables, editors, and analytics on top of it.
Delivered responsive layouts and improved keyboard accessibility across the portal.
-
Device metrics for ~3,000 signage players
A device metrics system for ~3,000 Android signage players using RabbitMQ, Node/Express, and TimescaleDB/Postgres, surfacing CPU, memory, disk, temperature, and playback data.
Supports fleet monitoring and production troubleshooting.
-
Observability migration: ELK → OpenSearch
Led the migration from Elasticsearch/Kibana to AWS hosted OpenSearch and Logstash, improving log ingestion, alerting, and storage lifecycle management.
Reduced infrastructure costs by ~10%.
-
RAG chatbot for a major U.S. mall operator
A RAG chatbot using Next.js, Python, OpenAI APIs, and Weaviate, with data refreshes automated through Lambda and EventBridge.
Retrieval quality and response accuracy evaluated before release.
-
A Next.js 15, React 19, and TypeScript research app that orchestrates parallel agents, live progress, typed session state, citations, retries, timeouts, and mock/live search adapters.
Deployed as a working AI research product at agents.dagos.to.
Outside of work, on my own time, I'm building Rally — an iOS app for fans who actually show up to live games: log every game, chat with fellow fans in the stadium, and build a record of every moment you were there.
skills
Lead Developer · mLOOP
April 2026 to Present
- Lead developer from initial architecture through production launch: matching engine, AI features, MCP server, deployments, and testing.
B.S. Computer Science · University of Connecticut
Storrs, CT
Toolbox
- Languages & Frameworks
- TypeScript, JavaScript, Python, SQL, C#, Java; Next.js, React, Vue 3, Fastify, .NET
- Data & Cloud
- PostgreSQL/TimescaleDB, MSSQL, Drizzle, RabbitMQ, AWS (Bedrock, EC2, RDS, S3, SES, Lambda), Docker, GitHub Actions, OpenSearch
- AI & Quality
- Claude Code, Codex, Cursor, MCP, RAG, OpenAI APIs, Playwright, Vitest, Sentry