# Gabriel Olarte Medrano > Software Engineer building MCP tooling and agent infrastructure at MCPJam. Based in Cochabamba, Bolivia. This document is a self-contained onboarding profile for an assistant researching or collaborating with Gabriel. It summarizes his own portfolio and CV; public evidence is identified with its specific scope below. ## Current role, education and availability Gabriel is a Software Engineer & Bolivia Site Lead — MCPJam. He works full-time on open-source MCP tooling and agent infrastructure, starting in 2026. Systems Engineering student at Universidad Privada Boliviana (UPB), Cochabamba. Studies began in 2024 and are ongoing. Location: Cochabamba, Bolivia. Timezone: GMT-4 (America/La_Paz). He is used to overlapping with EU and US working hours. Languages: Spanish native · English C1 · German B1. He is open to selected open-source, technical-community and speaking collaborations. Contact: olartgabo@gmail.com. Sources: [About and career](https://olartgabo.dev/#section), [Contact](https://olartgabo.dev/#notes), [CV](https://olartgabo.dev/cv/gabriel-olarte-medrano-cv.pdf). ## Working approach Competitive programming at IOI and ICPC shaped his approach to engineering: a solution must work at the scale and under the constraints it will meet in use. His portfolio connects that discipline to agent infrastructure, where confident incorrect answers can be a failure mode. His work spans cloud infrastructure, application delivery and local hardware. He builds AWS serverless systems for organizations, local inference on Raspberry Pi hardware, multi-agent decision support, computer vision and augmented reality. The About section emphasizes experience shipping for real users in Bolivia and designing for unreliable connections. Human oversight is part of the systems he describes. MosQuit stages a debate and simulation before a human signs off on an allocation; Javier requires approval before granting new tools at runtime. Olympus coordinates repositories, agent context and automation across machines. Source: [About](https://olartgabo.dev/#section) and the project case studies below. ## Career and organizations ### MCPJam Software Engineer & Bolivia Site Lead — 2026—. Full-time work on the open-source inspector and the surrounding Model Context Protocol ecosystem: tools developers use to build, test and debug MCP servers. Public source: https://github.com/MCPJam/inspector ### CloudNovaBo Co-Founder — 2025—. Enterprise AI solutions for corporate clients in Bolivia and the region, AWS serverless architecture, technical implementation and client delivery. Public source: https://cloudnovabo.com ### Meliorem Founder — 2025—. Enterprise AI automation using Lambda, SQS, API Gateway and the Claude API, including generative AI workflows and multi-agent architectures. ### QHub CTO, Innovation Department — 2026—. ### Clase360 Software Engineering & Technical Consulting — 2026. ### CochaTech Lead Developer & Training Coordinator — 2025—2026. Development and maintenance of the hackathon platform, plus technical workshops and participant training. The portfolio reports more than 1,100 participants served. Public source: https://cocha.tech ### Rebee CTO — 2025. AWS cloud infrastructure, technical architecture and backend development during product launch. ### Boring Soft Lead AI Developer — 2025. AI and fullstack development across products, including AI-powered features and workflow automation. ### Ensarta2 Full-Stack Developer — 2026. Public source: https://ensarta2.com ### Paleontological Museum AR Developer — 2025. Augmented-reality museum experiences. Role history source: [Portfolio organizations](https://olartgabo.dev/#section). Role descriptions: [CV](https://olartgabo.dev/cv/gabriel-olarte-medrano-cv.pdf). ## Community leadership and specialized training - 2025 — AWS Student Builder Group Bolivia: Founded the country's first official group at UPB. AWS Cloud Club Captain term began in 2025; 100+ active members. - 2026 — QHub Peru: Selected international student — quantum algorithms and programming frameworks. - 2026 — AWS Certified AI Practitioner: Proctored exam. The CV describes UPB Reactions, an interdisciplinary Neuroscience and AI club project involving fine-tuned model development, dataset preparation, parameter tuning and evaluation of academic reasoning behavior (2026–present). Source: [Career and training](https://olartgabo.dev/#section), [CV](https://olartgabo.dev/cv/gabriel-olarte-medrano-cv.pdf). ## Technical capabilities - LANGUAGES: C++ · Python · TypeScript · C# · Java. - AI / AGENTS: Claude API · MCP · Multi-agent · LangChain · Qwen. - CLOUD: AWS Lambda · SQS · API Gateway · EC2 · Serverless. - FRONTEND: React · React Native · Astro · Three.js · Tailwind. - VISION / HW: OpenCV · MediaPipe · ARCore · Raspberry Pi 5 · IR. Projects below identify where specific technologies were used. ## Selected projects and technology context ### MCPJam Role: Software Engineer & Bolivia Site Lead. Year: 2026. Open-source tooling for the Model Context Protocol — the inspector developers use to build, test and debug MCP servers. Stack: TypeScript, React, MCP, Node.js. The work: MCPJam builds developer tooling for the Model Context Protocol — the open standard that lets AI agents talk to tools and data. As a full-time Software Engineer and Bolivia Site Lead, I work across the open-source inspector and the agent infrastructure around it: the surface developers use to build, test and debug MCP servers. Why it matters: During my tenure, MCPJam's user base grew from approximately 30,000 to more than 50,000. MCP is becoming the wiring layer of the agent ecosystem, so the bugs I fix and features I ship land in other developers' daily workflow. Reported User base during tenure: 30K → 50K+. Reported Role: Software Engineer & Bolivia Site Lead. Case study: https://olartgabo.dev/work/mcpjam - [GitHub](https://github.com/MCPJam/inspector) ### Olympus Role: Creator. Year: 2026. A shared personal AI and automation architecture spanning an always-on assistant, an Arch Linux engineering workstation and a Raspberry Pi edge node. Stack: MCP, Arch Linux, Raspberry Pi 5, AI agents. The system: Olympus coordinates three machines through one shared architecture: Theseus, an always-on AI assistant and MCP orchestration node; Odysseus, an Arch Linux workstation for engineering and agent-assisted coding; and Telemachus, a Raspberry Pi 5 edge-computing node. The operating model: The machines share a repository, agent-specific context and cross-device automation. Instead of treating assistants, development environments and hardware as separate experiments, Olympus makes them parts of one system that can coordinate work across devices. Reported Nodes: 3 machines. Reported Architecture: Shared agent context. Case study: https://olartgabo.dev/work/olympus ### MosQuit Role: Creator. Year: 2026. Four specialized agents debate mosquito-control resource allocation for public-health departments, with satellite intelligence and 7-day outbreak simulation. Stack: Claude API, Multi-agent, Satellite data. The problem: Public-health departments allocate mosquito-control resources under uncertainty: incomplete field data, shifting weather, competing priorities. A single model gives a single opinion. MosQuit stages a structured debate instead. The system: Four specialized agents — epidemiology, environment, logistics and budget — argue over the same decision, grounded in satellite environmental data. A 7-day outbreak simulation stress-tests each proposed allocation before a human signs off. The result: Selected in the top 500 of more than 25,000 proposals in Anthropic's global hackathon. Reported Anthropic global: Top 500 / 25,000. Reported Forecast window: 7-day simulation. Case study: https://olartgabo.dev/work/mosquit - [Demo](https://mosquitgabo.vercel.app/) - [GitHub](https://github.com/olartgabo/MosQuit) ### Javier Role: Creator. Year: 2026. A personal assistant running entirely on a Raspberry Pi 5 — local quantized LLM, self-growing MCP architecture, infrared eye tracking, OS and smart-home control, 24/7 via Telegram. Stack: Raspberry Pi 5, Qwen (local), MCP, OpenCV. The premise: No cloud, no subscription, no data leaving the room: a personal AI assistant that lives on a Raspberry Pi 5 on my desk, running a quantized Qwen model locally. The architecture: A self-growing MCP architecture: Javier can be granted new tools at runtime, each one gated behind human approval. An infrared camera adds eye tracking — look left and the active window snaps left. It controls the OS, the smart home, and answers on Telegram around the clock. Reported Inference: 100% local. Reported Availability: 24/7. Case study: https://olartgabo.dev/work/javier ### XploRA Role: Founder. Year: 2024. Augmented-reality tourism with monument recognition and museum experiences, in production with real users across Bolivia. Stack: Unity, C#, Google ARCore. The product: Point a phone at a monument and XploRA recognizes it, overlaying its history in augmented reality. The same engine powers museum experiences — including work for the Paleontological Museum. In production: 100–500 active users in production, built in Unity with Google ARCore. Backed by the GAMT startup incubator with a 21,000 Bs award. Reported Active users: 100–500 in production. Reported GAMT incubator award: 21,000 Bs. Case study: https://olartgabo.dev/work/xplora ## Additional public engineering work - [design-dna-mcp](https://github.com/olartgabo/design-dna-mcp) (TypeScript): Extract the design DNA of any website — an MCP server that turns sites into a searchable library of reusable UI patterns. - [boxMouse](https://github.com/olartgabo/boxMouse) (Python): Xbox 360 controller as a Windows mouse, Claude Code pad and komorebi remote. Zero dependencies — just ctypes. - [HandTracking](https://github.com/olartgabo/HandTracking) (Python): Computer-vision system for gesture-based PC control, built on MediaPipe and OpenCV. - [windows-dotfiles](https://github.com/olartgabo/windows-dotfiles) (PowerShell): Tokyo Night Windows dev rice: komorebi, whkd, YASB, LazyVim and PowerShell 7, with a sync script to move it between machines. - [mcpUPB](https://github.com/olartgabo/mcpUPB) (TypeScript): Model Context Protocol work for Universidad Privada Boliviana. - [awscloudclubUPB](https://github.com/olartgabo/awscloudclubUPB) (TypeScript): Site for the AWS Student Builder Group chapter at UPB. - [Authored MCPJam pull requests](https://github.com/MCPJam/inspector/pulls?q=is%3Apr+author%3Aolartgabo): Public contribution history with implementation and review discussions. ## Credentials and evidence The following achievements are reported in Gabriel's portfolio. Linked public evidence supports only the scope stated beside it; other references are available by email. - 2026 — Anthropic global hackathon — MosQuit selected worldwide. TOP 500; OF 25,000 PROPOSALS. - 2025 — International Olympiad in Informatics — Bolivia national representative. IOI 2025; TOP 8 NATIONAL. - 2024 — ICPC Latin America regional final, first university participation. ICPC LATAM; 2ND BEST IN UPB HISTORY. - 2025 — Founded Bolivia's first official AWS Student Builder Group at UPB; 100+ active members. AWS; STUDENT COMMUNITY LEADERSHIP. Evidence: [AWS Community Day speaker listing](https://day.awscommunity.org.bo/es#speakers). Confirms the UPB Cochabamba group leadership role. Founding and membership references available on request. - 2025 — CochaTech — platform for Bolivia's largest hackathon. 1,100+; PARTICIPANTS SERVED. - 2025 — National GameJam — selected to represent Bolivia abroad. 1ST PLACE; NATIONAL GAMEJAM. - 2026 — Nodi Scholar recognition — emerging LATAM technology talent. $1,000; NODI SCHOLAR AWARD. - 2024 — GAMT startup incubator award for XploRA. 21,000 Bs; GAMT INCUBATOR. - 2025–26 — AWS certification scholarships granted directly by Amazon. AWS; CERTIFIED AI PRACTITIONER. - 2023 — UCB Cybersecurity Hackathon: 1st place. - 2026 — Authetical: 2nd place at CochaTech. Accepted by BMSC as a real-world use case. MCPJam's reported 30K → 50K+ user-base growth describes the period of his tenure; it is not a measurement of growth caused solely by Gabriel. XploRA's 100–500 active users, competition placements and awards are self-reported portfolio figures. Public repository links demonstrate available work, not independent validation of every outcome claim. ## Common questions ### Who is Gabriel Olarte Medrano? Gabriel Olarte Medrano is a Software Engineer & Bolivia Site Lead — MCPJam, based in Cochabamba, Bolivia. He builds Model Context Protocol (MCP) tooling, AI agent systems and cloud infrastructure. ### What has Gabriel built? His selected work includes MCPJam's open-source MCP inspector, Olympus distributed AI infrastructure, MosQuit multi-agent decision support, Javier's Raspberry Pi assistant, and XploRA augmented-reality tourism experiences. ### What technologies does Gabriel work with? TypeScript, React, Python, C++, MCP, Claude API, AWS serverless services, Raspberry Pi, OpenCV, Unity and ARCore. Each case study lists the technologies used in that project. ### How can I collaborate with Gabriel? Gabriel works full-time at MCPJam and is open to selected open-source, technical-community and speaking collaborations. Contact olartgabo@gmail.com. His timezone is GMT-4, with overlap across EU and US working hours. ## Sources and contact - [Quick brief](https://olartgabo.dev/hire) - [Credentials](https://olartgabo.dev/#proof) - [CV](https://olartgabo.dev/cv/gabriel-olarte-medrano-cv.pdf) - [GitHub](https://github.com/olartgabo) - [LinkedIn](https://linkedin.com/in/gabriel-olarte-medrano-741aa331a) - [Email](mailto:olartgabo@gmail.com) - [Human-readable agent onboarding](https://olartgabo.dev/for-agents) - [Markdown version](https://olartgabo.dev/for-agents/index.md)