About me

I am a Systems-oriented Mobile Architect and AI Infrastructure Researcher with over 10 years of experience building software that scales. I'm currently working on closing the reliability and performance gap between high-throughput backends and world-class mobile user experiences.

Current Mission: ZEA Platform (Oct 2024 – Present)

Founder & Lead AI Architect

At ZEA, I'm building an Enterprise AI Infrastructure Platform that combines three unique capabilities: AI Gateway with FX optimization (like Stripe for AI payments), Workflow Orchestration with Human-in-the-Loop (like Temporal + mobile approvals), and a unified multi-tenant SaaS platform. I identified that while LLMs are powerful, deploying resilient, agentic workflows with human oversight is still broken.

  • Cerebelum (Workflow Engine): I architected a next-generation orchestration engine using Elixir/OTP. It merges the stateful graph management of LangGraph with the workflow durability of Temporal.io, adding a unique Human-in-the-Loop capability — workflows pause for mobile approvals with biometric authentication, then resume automatically. The infrastructure handles 640K events per second using event-sourcing, featuring "time-travel" debugging and long-running process hibernation.
  • Cortex (AI Gateway): Developed a resilient multi-provider AI Gateway supporting OpenAI, Anthropic, Gemini, Groq, Cohere, xAI, and Ollama. Features sub-millisecond failover, intelligent load balancing, and FX optimization that saves LATAM startups 1.4-1.6% on currency exchange when paying AI providers in USD — creating a network effects moat as volume scales.
  • Developer Experience: Designed Python SDK and Kotlin Multiplatform SDK (iOS/Android) with DX-first approach, bridging complex Elixir backends with simple, ergonomic APIs that minimize cognitive load for developers building agentic workflows.

The Scaling Years: MOBDEV (2018 – 2024)

Principal Software Engineer (Mobile)

I joined MOBDEV when the team was just 10 people and helped scale the organization to over 100 members. As Principal Engineer of the Android unit (15+ devs), I navigated the complexities of hyper-growth while maintaining technical excellence.

  • Hybrid Architecture: Led the migration to a high-traffic React Native stack. When the bridge became a bottleneck, I built custom Native Modules (Swift/Kotlin) to eliminate TTI issues and ensure instantaneous responsiveness.
  • Multiplatform Pioneer: Spearheaded the adoption of Kotlin & Compose Multiplatform (KMP/CMP), reducing maintenance costs by 30% and unifying the engineering culture.
  • Leadership: Established a technical career ladder that promoted 5 junior engineers to Senior/Lead roles within 24 months.

Foundations: R&D and Fintech

My approach to engineering is rooted in mathematical rigor and low-level optimization:

  • Medical R&D (PUC Chile): I migrated cardiovascular software for 3D blood flow quantification using Finite Element Methods (FEM). I engineered the integration layer between low-level Objective-C libraries and Python, reducing image processing time by 50%.
  • Fintech (Multicaja): Modernized mission-critical financial apps, implementing TDD and Clean Architecture.
  • Entrepreneurship: As Co-Founder of Loadingplay, I managed the full lifecycle of 12+ MVPs, mastering the art of rapid market validation.

Academic & Research

I am currently a Master of Science in Computer Science candidate at Universidad de Santiago de Chile. My thesis focuses on implementing OTP-style concurrency primitives in Rust for memory-safe, high-throughput AI agent orchestration.

I'm @chinostroza on GitHub, you can find me on LinkedIn, and contact me at c@zea.cl.

Frutillar, Chile • +56992246145 • English: Full Professional Proficiency

Experience Highlights

Oct 2024 – Present

ZEA Platform

Founder & Lead AI Architect

Cerebelum: Architected a next-generation orchestration engine in Elixir/OTP, bridging the gap between LangGraph's stateful graph management and Temporal.io's workflow durability. Designed the system to handle high-throughput event-sourcing for deterministic AI agent execution, featuring long-running process hibernation and "time-travel" debugging capabilities.

Cortex: Developed a resilient AI Gateway featuring sub-millisecond failover and intelligent routing across OpenAI, Anthropic, and local LLMs, reducing API costs through optimized prompt caching and token rotation.

Systems Integration: Adopted a DX-first approach in designing Python SDK and Kotlin Multiplatform SDK that bridge mobile edge devices with complex Elixir backends, successfully minimizing workflow cognitive load and reducing implementation friction.

Apr 2018 – Sep 2024

MOBDEV

Principal Software Engineer (Mobile)

Organizational Impact: Played a pivotal role in scaling the engineering organization from 10 to 100+ members while directing the Android unit (15+ devs).

Architecture Strategy: Led the technical migration of high-traffic platforms to a hybrid React Native stack, building custom Native Modules (Swift/Kotlin) that solved TTI bottlenecks and significantly enhanced overall application responsiveness.

Scale: Spearheaded the adoption of Kotlin & Compose Multiplatform (KMP/CMP) to share both business logic and UI components across platforms. This initiative resulted in a 30% reduction in maintenance costs and a highly unified engineering culture for iOS and Android.

Mentorship: Established a technical career ladder that promoted 5 junior engineers to Senior/Lead roles within 24 months, increasing team retention by 20%.

May 2017 – Oct 2018

Multicaja S.A.

Android Engineer (Fintech)

Modernized a mission-critical financial app, implementing TDD and Clean Architecture. Achieved 99.9% crash-free sessions and reduced critical bug reports by 60% within the first 6 months.

Aug 2015 – Sep 2016

Pontificia Universidad Católica de Chile

Software Engineer (R&D)

Advanced Medical Imaging: Developed cardiovascular software for 3D blood flow quantification and structural measurements using Finite Element Methods (FEM).

Systems Interoperability: Engineered a high-performance integration layer between low-level Objective-C libraries and Python, exposing complex C-based image quantification algorithms to researchers and scientists for advanced data analysis.

Optimization: Optimized volumetric data processing pipelines, reducing image quantification and processing time by 50% while maintaining precision for clinical research.

Jan 2012 – Dec 2014

Loadingplay

Co-Founder (Startup Factory)

Directed the full tech lifecycle for a portfolio of digital products. Managed the delivery of 12+ MVPs, leading to successful market validation.

"I build mobile products that feel instantaneous. My approach combines deep native knowledge with modern distributed systems principles, ensuring that architecture scales as fast as the user base."