Justin Rhinesmith

First engineering hire at Boon (getboon.ai), the AI intelligence layer for construction — agents that run preconstruction workflows from blueprint to bid. Built core platform systems across two product generations (freight automation → construction estimation), partnering directly with the CEO and early customers.

• Architected the workflow engine (Rails, Sidekiq, PostgreSQL) for LLM-powered document ingestion — PDF parsing, structured extraction against enforced schemas, retries, and human-in-the-loop review — used across both freight automation and construction estimation, including semantic mapping from customer-specific catalogs into canonical bid/takeoff data.
• Core engineer on the AI Takeoff platform: collaborative multi-user annotation with an offline mutation queue and server-side conflict reconciliation, a server-side PDF pipeline handling 300+ page drawing sets, and the V1→V2 canvas migration — adding deterministic canvas test hooks and Playwright fixtures that drove the canvas in CI and caught migration regressions.
• Built the agent-native engineering harness — CI eval gates, acceptance-criteria-driven verification, architectural fitness tests, and autonomous maintenance workflows — that let the team ship at high volume (1,400+ reviewed PRs in 12 months) without loosening regression coverage or release-health gates.
• Own production reliability and observability: OpenTelemetry (Grafana, Loki, Tempo, Prometheus), release-health-gated progressive rollouts, and a chaos-engineering harness that surfaced and fixed real resilience bugs (unbounded queries, error-classification faults, third-party-outage worker hangs) before they reached customers.
• Shipped self-serve monetization end to end: Stripe subscriptions and token top-up billing, plus automated per-org trial provisioning for the Boon Agent product.
• Hardened the LLM gateway: model routing, provider circuit breaker with mid-stream fallback, and Bedrock prompt caching; cost work eliminated the single largest line item of third-party model spend.

Joined pre-IPO and helped scale through the December 2020 public offering and pandemic-era hypergrowth. Tech lead across Category Expansion (Pharmacy → Convenience & Health).

• Led the HIPAA-compliant DoorDash Pharmacy pilot: isolated identity and data boundaries, restricted access paths, and auditability, coordinated across engineering, product, and legal.
• Architected Inventory Service and co-designed the Restriction Rules Service for eligibility and compliance constraints used across Alcohol/Health.
• Drove HSA/FSA eligibility & reimbursement features; coordinated cross-functional delivery and scoping across Category Expansion.
• Ran Category Expansion design reviews and mentored engineers through architecture reviews.

Tech lead on the Caviar engineering team, driving scalability, data privacy, and post-acquisition integration work between Caviar and DoorDash.

• Led the Caviar-to-DoorDash user migration for millions of accounts: owned the runbook — data migration, identity reconciliation, backfill validation, and rollback plan — for a low-downtime acquisition cutover.
• Helped implement CCPA compliance features across customer data services, ensuring secure data export and deletion flows in alignment with privacy regulations.
• Diagnosed and resolved scaling bottlenecks across core Rails and Postgres systems, improving throughput and API reliability during peak ordering windows.

Promoted from Lead Dev; managed Android & React products backed by Rails.

• Led a 7-person team; introduced CI/CD and versioned APIs; scaled Rails/Postgres/Redis.

Built and launched a subscription fitness platform — Swift iOS app featured in Apple's Best New Apps and Health & Fitness, backed by a Rails API I architected end to end.

• Designed a versioned JSON API (OAuth2) powering the iOS app: streaming workout video, social following, and scheduled group workouts.
• Shipped dual-rail subscription billing (Stripe + Apple in-app purchase receipt validation) with gifts, vouchers, and referral credits.
• Integrated Fitbit, Jawbone, and Spotify; Elasticsearch-backed discovery; Sidekiq/Redis background processing — with RSpec + Cucumber coverage and CI.

Led engineering on live-service mobile/social games across iOS, Facebook, and Google platforms.

• Led 10 engineers; kept sub-250ms latency while serving 180k+ RPM; scaled Rails/Node backends to 1M+ MAU.
• Built high-throughput game backends (Rails, MySQL, Redis) with cross-platform in-app-purchase validation, tournament/leaderboard systems, and a StatsD-instrumented live-ops and analytics pipeline.

Multiple Rails apps for content/document delivery; introduced testing, code reviews, CI/CD.