Ulises Rosas

Co-founded a boutique remote software studio delivering full-stack, mobile, and web products for U.S. and LATAM startups. Hands-on engineer across the full SDLC while leading distributed, cross-functional teams.

• Led delivery of 5 end-to-end client products, 4 on time and within budget and scope.
• Shortened typical feature release cycles from 1–3 months to 3–4 weeks by introducing mentorship, code reviews, and shared architectures.
• Cut production defects 50% and release cycle time 20% across 3 client products in 12 months, standardizing error tracking, CI/CD, and automated deployment tooling.
• Increased client revenue and active users while lowering infrastructure costs through CI/CD, automated releases, and in-house analytics.
• Improved user retention by building in-app messaging and supporting marketing campaign tooling.

Fitness platform: web admin, push notification service, and iOS/Android app.

Stack: React Native, Typescript, Firebase, Node.js, n8n, Bluetooth, Sentry, CI/CD

• Owned the iOS/Android app and Bluetooth device data capture end to end, from requirements to store release.
• Reduced end-to-end release time from 8 weeks to 3 weeks across 14 releases in 12 months.
• Grew monthly active users from 4,400 to 5,000 (+10%) with AI-personalized in-app recommendations.
• Built a model-agnostic recommendation layer in n8n integrating Gemini 2.5, GPT-4.1-mini, and Deepseek.
• Introduced Sentry plus stronger QA standards, driving error detection and resolution to near immediate.

Companion app tied to a host device.

Stack: React Native, MQTT, GitHub Actions, E2E Testing, Sentry

• Reduced crashes 50% and improved MQTT reliability 65%.
• Handled high-throughput MQTT traffic (10–15 packets/sec) with background processing and queue-based delivery.
• Built 3 automated tests plus 1 end-to-end suite, cutting total test time 40%.
• Automated publishing with GitHub Actions, reducing release time from 1 hr to 15 min.
• Shipped to the Apple App Store and Google Play; serves 50+ active users.

Patient tracking MVP for clinics and medical centers.

Stack: Next.js, AWS Amplify, Cognito, DynamoDB, S3, CI/CD, React, Serverless

• Reduced deployment time from 25 min to 5 min by building CI/CD pipelines across dev, staging, and production.
• Architected the AWS stack: Amplify, Cognito for auth, DynamoDB for data, S3 for media.
• Raised automated test coverage from 40% to 90%, holding incidents at 15% with zero rollbacks.
• Delivered an MVP handling 500 daily requests with Amplify auto scaling.
• Served 100 concurrent users at launch, architected to scale toward 5,000+.

Real-time patient vital signs monitor for a clinic (200+ rooms/staff).

Stack: Next.js, React, Node.js, Docker, GitHub Actions

• Reduced page load times from 1.5s to 1s using Next.js SSR, lazy loading, and CDN-served static content.
• Built the vital signs rendering components and push notification receiver for floor staff alerts.
• Automated Docker + GitHub Actions pipelines, cutting deploys from 20 min to 5 min.