
Pau Gallardo Cordon
Verified Expert in Engineering
Python Developer
Andorra, Andorra
Toptal member since September 23, 2026
Across air transport, AI, and blockchain, Pau has spent eight years building full-stack and AI systems. His toolkit centers on TypeScript, Python, and PostgreSQL. While at SITA, he took full technical ownership of Mission Watch Multi-platform, carrying it from an empty repository to production and cutting CPU consumption by 30%.
Portfolio
Experience
- Python - 10 years
- Fastify - 8 years
- TypeScript - 8 years
- PostgreSQL - 8 years
- Next.js - 5 years
- AI Architecture - 4 years
- Pgvector - 3 years
- LangGraph - 3 years
Preferred Environment
LangChain, LangGraph, LangSmith, Node.js, TypeScript, Next.js, NestJS, Fastify, PostgreSQL
The most amazing...
...product I've built is Nubion AI, which reached 7,000 active users and a €5 million valuation in under eight months as a fully bootstrapped platform.
Work Experience
Senior Full-stack Engineer
SITA
- Took Mission Watch Multi-platform from an empty repository to production, assuming full technical ownership six months before launch and delivering the release successfully.
- Improved system performance by reducing CPU consumption by 30% and map element load by 10%, including a major redesign of the map-rendering architecture using Canvas, MapLibre, and Konva.
- Served as technical implementation lead, hiring, onboarding, and training new engineers while defining the architecture and engineering practices used by the team.
- Built a reusable AI skillset powered by custom MCP integrations and internal tooling that enabled engineering and QA teams to automate recurring development, validation, and operational workflows across the aviation platform.
- Developed Jira automation skills that allowed engineers to create, update, transition, and organize tickets programmatically through JQL-driven workflows, reducing dependency on manual Jira interactions and standardizing issue-management processes.
- Implemented an AI-assisted UI validation skill connected to real aviation back-end environments and production-like data, enabling manual QA engineers to automate portions of regression testing and application behavior against live service responses.
- Created automated pull-request review workflows using Claude hooks and custom validation rules to inspect code changes, verify contextual and technical requirements, and surface potential inconsistencies before human code review.
- Designed service-health validation skills that checked back-end APIs and 3rd-party integrations for availability, validated dependencies between services, and verified that critical system components were communicating correctly across the platform.
- Built LangGraph workflows to orchestrate multi-step engineering tasks with stateful execution, conditional branching, retries, and validation before progressing to the next stage.
- Designed multi-agent LangGraph flows that separated planning, implementation, review, and validation responsibilities, improving traceability and reducing unsafe autonomous actions.
Software Engineer (Contractor)
Caamaño Asia
- Automated the AutoCAD review that determines what gets manufactured for each store, taking the review team from 13 product owners down to five.
- Automated the extraction and checking of manufacturing requirements straight from their AutoCAD drawings, replacing most of the manual review pass.
- Led the integration with the Chinese team, overcoming language and cultural barriers.
- Developed a multi-provider AI skill for product owners that triggered AutoCAD automation scripts, calculated manufacturing quantities automatically, and organized the resulting documents into predefined project folders.
Founder and CTO
Nubion AI
- Founded the company and led it as CTO: built a team of five engineers and took the product to 7,000 active users in under eight months, reaching a €5 million valuation fully bootstrapped.
- Built the product on LangGraph from its v0 releases, contributing back alongside the community as the framework took shape.
- Ran the data layer on PostgreSQL with pgvector and instrumented the agents with LangSmith for tracing and evaluation.
- Built Nubion’s own automated customer support system in-house.
- Built the marketing measurement and optimization stack in-house, so growth never depended on an external agency.
- Developed a Claude Skill that helped support teams resolve customers' technical questions through guided troubleshooting, reducing engineering escalations, and later migrated it into a reusable plugin.
- Built and scaled an AI platform from greenfield to 7,000+ active users, owning architecture, back-end services, front-end delivery, deployment, and production reliability.
- Designed production RAG pipelines combining embeddings, vector search, structured retrieval, and contextual grounding to improve answer relevance across user workflows.
- Implemented agentic AI workflows with tool calling, validation steps, and controlled execution paths to automate complex multi-step user tasks reliably.
- Established observability and evaluation practices for AI features, tracking failures, retrieval quality, latency, and model behavior to support continuous production improvement.
Founder and Software Engineer
Alos Innovation
- Founded the company in 2021 and grew it to 20+ clients and roughly €500,000 in annual revenue, bootstrapped, leading teams of up to 13 people across engagements.
- Delivered for clients across financial services, healthcare, energy, fuel distribution, retail, and logistics.
- Built a system that operates as a software factory and as the vehicle through which I take on my own contract engagements.
- Built a Claude Skill for creating and updating Jira issues through MCP and JQL, standardizing ticket structure, duplicate checks, project resolution, and release-epic linking.
- Designed an SDD-based engineering skillset that guided agents and developers from requirements through specification, implementation, validation, and final technical judgment.
- Implemented structured judgment phases before and after development to challenge assumptions, validate architectural decisions, and catch inconsistencies before code reached review.
- Created reusable Claude Skills that converted product requirements into implementation plans, Jira tasks, acceptance criteria, and validation steps before development started.
- Built an agent workflow that combined Jira MCP, JQL, custom tools, and project context so multiple AI agents could independently pick up and execute well-defined engineering tasks.
- Standardized AI-assisted development around Specification-Driven Development, requiring explicit requirements, technical design, implementation planning, and validation before completion.
- Established an AI-native software delivery process across the company, combining Claude Skills, MCP tools, SDD, Jira automation, and agent-based code review into a repeatable engineering workflow.
Senior Software Engineer (Contract)
Bordr LLC
- Delivered the full contracted scope across the front end and back end on a platform handling around 10,000 requests a month, and owned the CI/CD pipeline that the entire team shipped through.
- Built tracking as a semi-automated process around a human-in-the-loop stage that could not be automated away.
- Cut overall churn by 0.3% by automating around that stage.
- Built and shipped MVP features end to end with React, Node.js, NestJS, Express, GraphQL, Prisma, and PostgreSQL.
- Defined the quality gates the pipeline enforced (SonarQube, unit tests, E2E), giving the team traceable per-developer code quality against Bordr’s standards.
- Drove architectural decisions for the end-to-end flow alongside the founders.
- Onboarded and mentored a junior developer into the codebase and architecture.
- Improved document-processing workflows across the platform, reducing manual handling and making customer case files easier to track, validate, and manage throughout their lifecycle.
- Strengthened platform stability by identifying and resolving recurring back-end and integration failures, improving reliability across critical customer-facing workflows.
- Optimized database queries and back-end processing for document-heavy workflows, reducing unnecessary load and improving response times across frequently used platform operations.
Full-stack Developer
Nerazim
- Carried both the product and the blockchain layer, and owned the hardest piece of it: the shader pipeline that merges separate robot part NFTs into a single newly minted NFT.
- Co-wrote the game’s smart contracts and the 3D shaders.
- Made the 3D merge work end-to-end: shaders combining the source models, and a custom binary format designed specifically to make the merged geometry come out correct.
- Held the whole pipeline under 30 seconds, with very heavy assets and against real WebGL performance limits.
- Optimized the WebGL rendering pipeline to handle complex merged 3D assets within browser performance limits while keeping the full processing flow under 30 seconds.
- Designed and implemented asset-processing workflows that transformed multiple 3D models into game-ready geometry while preserving materials, metadata, and visual consistency.
- Integrated blockchain logic with the game layer, connecting smart contracts, asset ownership, and in-game state across front-end and back-end systems.
- Built reusable rendering and geometry-processing components that reduced duplicated shader logic and simplified the addition of new NFT-based game assets.
- Improved client-side performance for heavy 3D scenes by reducing unnecessary geometry, draw calls, and processing overhead across the WebGL pipeline.
- Implemented validation around blockchain transactions and asset state to keep on-chain ownership synchronized with the game’s internal data model.
Front-end Developer (Freelance)
Ambulatory EHR, Hôpital du Valais
- Joined an ambulatory electronic health record project months from go-live and gave an under-pressure team the capacity it needed to hit its release date.
- Ramped on an unfamiliar Angular system fast enough to know where a fix would actually move the release, not writing the fixes themselves.
- Delivered on a very tight schedule with almost no time for onboarding.
- Stabilized critical application flows ahead of go-live by troubleshooting complex front-end issues across a large enterprise Angular platform.
- Collaborated closely with product, QA, and engineering teams to prioritize defects and ensure critical functionality was ready for production release.
- Reduced delivery risk during the final stages of the project by quickly taking ownership of unfamiliar modules and resolving high-priority technical issues.
- Improved maintainability of existing Angular components through targeted refactoring while avoiding unnecessary changes during a release-critical period.
- Supported the successful transition toward production by balancing feature delivery, defect resolution, and regression prevention in a healthcare environment.
Process Optimization Engineer
Giesecke+Devrient
- Improved the image-recognition algorithm on an automatic card-inspection machine to cut false positives, dropping scrap by 10%, lifting the machine’s OEE substantially, and improving time and cost by over 25%. The project became my final-year thesis.
- Optimized detection in Python to remove most of that rework caused by false positives, sending good cards into a second inspection pass.
- Graduated with honors with the project implemented in this company.
- Designed and validated image-processing improvements using production data, increasing detection reliability while reducing unnecessary rejection of correctly manufactured cards.
- Deployed the final computer vision solution into the company’s production environment, turning an academic engineering project into a real industrial automation system.
- Reduced unnecessary inspection rework by improving classification accuracy and tuning detection thresholds against real manufacturing defects and edge cases.
- Analyzed machine-performance and inspection data to identify the main causes of false positives and prioritize the algorithmic improvements with the greatest operational impact.
Experience
Nubion AI
https://www.nubionai.com/enMy involvement has gone beyond the technical side, helping to shape both the product itself and how it is positioned, communicated, and brought to market. This end-to-end involvement has given me hands-on experience in building and scaling a real-world AI product, balancing technical decisions with user needs, growth, and business objectives.
SITA Mission Watch Multi-platform
https://www.sita.aeroI also worked extensively on the modular Angular architecture, integrating shared UI, map, and panel components; maintaining dependencies; upgrading Angular versions; and resolving complex build and native compatibility issues across web and desktop environments, including Apple Silicon. The role involved working across the full application lifecycle, from low-level technical troubleshooting and build tooling to core product features and architectural improvements.
Nubion LLC
https://www.nubionai.com/enEducation
Bachelor's Degree in Industrial Engineering
ETSEIB, Universitat Politècnica de Catalunya - Barcelona, Spain
Skills
Libraries/APIs
React Query, React, Node.js, WebGL, NgRx, Konva, REST APIs, Claude API, React Redux, Pydantic, Three.js, GraphQL API
Tools
Prisma, SonarQube, Prettier, Canvas, Tabnine, AI Prompts, ChatGPT, Claude, Claude Code, ComfyUI, MATLAB, SOLIDWORKS, Capacitor, AutoCAD
Languages
TypeScript, JavaScript, HTML, Python, AutoLISP, GraphQL, Solidity, C++, Rust, F#
Frameworks
Express.js, LangGraph, Fastify, Next.js, NestJS, Angular, React Native, Redux, Tailwind CSS, Jest, Cypress, Jasmine, Electron, Django
Paradigms
Agile Software Development, Database Design, Model Context Protocol (MCP), Real-time Systems, REST, DevOps, Desktop App Development, Requirements Analysis
Platforms
LangSmith, Docker, Harness, Amazon Web Services (AWS), AWS Lambda, Azure, Ethereum
Storage
PostgreSQL, MySQL, MongoDB, Amazon DynamoDB, Relational Databases, Elasticsearch, Amazon S3 (AWS S3), Datadog, Data Pipelines
Other
AI Architecture, Image Recognition, OpenAI, Anthropic, Maplibre, Maps, Azure Maps, Pgvector, Express, CI/CD Pipelines, MapLibre, LangChain, ESLint, Product Development, Software Consulting, High Code Quality, API Design, Cross-functional Collaboration, Greenfield Development, Legacy System Integration, Vector Databases, Front-end Architecture, Application State Management, Team Mentoring, AI Tooling Integration, MCP Server Development, Delivery, Canvas 3D, APIs, Coding, Artificial Intelligence (AI), Document Parsing, Architecture, Web Applications, AI Tools, Full-stack, Front-end Development, Back-end Development, User Roles, Workflows, Back-end, Prompt Engineering, Large Language Models (LLMs), AI Agents, AI Agent Orchestration, Agentic AI, AI Workflow, ChatGPT Prompts, Agentic Coding, AI Engineering, Cursor AI, Stakeholder Management, Clustering, Binary Data Processing, Hiring & Onboarding, Security-Constrained AI Systems, CSV, FastAPI, Pipelines, Desktop App, WebSockets, Optimization, Industry 4.0, WebAssembly (Wasm), Ethers.js, Shaders, WebAssembly, Industrial Engineering, Process Optimization, Manufacturing Processes, Industrial Automation, Production Systems, Continuous Improvement, Operational Efficiency Improvement, OEE Optimization, Quality Control (QC), Computer Vision, Data Analysis, Engineering Problem Solving, Systems Thinking, Process Analysis, Root-cause Analysis (RCA), Performance Optimization, Applied Mathematics, Engineering Design, Technical Project Management, R&D, Digital Transformation, Automation Technologies, Industrial Digitalization, Software Architecture, Full-stack Development, Technical Leadership, Engineering Management, Team Leadership, Team Building, Technical Mentoring, Project Delivery, Client Management, Solution Design, System Design, Technical Consulting, Process Automation, Testing Strategy, Technical Decision-Making, Memory Optimization, Microsoft Azure
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring