Mauricio – Node.js, Typescript, SQL

A set of client-facing web platforms running on AWS, delivered by a software consultancy for its client. The platforms share third-party user management with single sign-on (SSO) across products. The client's in-house development team is scaling up, so new work has to integrate cleanly with their codebase and standards.

• Led end-to-end feature development, from technical design and architecture to scoping and delivery.
• Implemented SSO and Auth0 from scratch, giving the client secure third-party user management across its platforms.
• Modernized the AWS infrastructure, tuning performance to reduce latency and optimizing tooling to lower operational costs.
• Worked closely with the client's development team during its staffing scale-up to keep integration seamless and maintain standardized documentation with developer READMEs and operational guides.
• Used AI agents (Cursor, Claude) for spec-driven development while following the client's coding standards.

A web gallery where event attendees and organizers browse and share event photos, with images sourced from Google Drive. Large events produce large photo sets, and serving every image directly from Drive results in many repeated upstream requests. The app runs on a serverless architecture, so it scales with uneven, event-driven traffic while keeping hosting costs low.

• Owned the gallery end-to-end, from requirements gathering and database design to the serverless back end and deployment.
• Built a Cloudflare R2 cache in front of Google Drive images, cutting repeated upstream requests to Drive.
• Implemented a sliding-window image-loading approach in React to keep large galleries responsive.
• Selected and tuned AWS serverless services (Lambda, API Gateway, Cognito, DynamoDB, Amplify) to keep cloud costs within the client's budget.
• Used AI agents (Cursor, Claude) for spec-driven development and test/edge-case generation, reviewing generated code for performance, security, and project conventions.

An enterprise AI demand-forecasting product for supply-chain businesses that need to predict demand across large product and location hierarchies. Planners use it to run what-if simulations, changing inputs and seeing the impact on the forecast before committing to a full prediction run. Running full model inference over hundreds of thousands of hierarchy points took hours, so interactive scenario planning wasn't practical. The product is built on a proprietary JavaScript/Java/Python AI platform with LLM, ML, and NLP capabilities.

• Led development of the demand-forecasting pilot, redesigning the what-if simulation engine to cut run time from 4–5 hours to ~10 seconds while keeping full inference for the final prediction; delivered within a two-week deadline, and the pilot won the client contract.
• Replaced repeated full-model inference with hierarchical caching, weighted aggregation/disaggregation and post-order traversal, accepting a ~5–6% accuracy trade-off to make real-time scenario planning possible.
• Aligned stakeholders on the approach by building demos of both options, making the accuracy vs. performance and CPU-cost trade-off visible.
• Led cross-functional teams of 8+ engineers across parallel feature deliveries, mentoring junior developers and running code reviews.
• Designed, built and deployed AI-based features using proprietary LLMs, ML models and NLP with Python on AWS EC2, plus full-stack features in React and TypeScript.
• Set engineering documentation standards and Agile practices across multiple teams.

Mission-critical core banking applications for a banking client, supporting day-to-day banking operations and back-office processing. The system is built on a Java Spring Boot microservices architecture, with scheduled batch jobs running core banking operational processes.

• Led a development team of 5+ engineers delivering features for mission-critical banking applications on a Java Spring Boot microservices architecture.
• Ran sprint planning, backlog grooming, and daily standups within a strict Scrum framework.
• Owned technical design documentation, feature analysis, and end-to-end delivery pipelines for the team.

Core banking applications for a banking client, handling day-to-day banking operations and back-office processing. The platform runs on Java Spring Boot microservices backed by SQL databases, with scheduled batch jobs (BEAM processes and JSR batch executions) that keep core banking operations running.

• Built features for core banking applications using Java Spring Boot microservices and SQL databases.
• Coordinated and implemented BEAM processes and JSR batch job executions as part of core banking operational support.
• Took part in technical design and feature analysis for new banking functionality; promoted to Technical Lead in April 2022.

Custom web applications and microservices built for business clients of a software development company, taken from initial concept through production release. The products expose RESTful APIs and web services, backed by relational and non-relational databases, designed to remain fast and responsive as usage grows.

• Developed and deployed full-stack web applications and microservices using Node.js, NestJS, C#, MySQL and MongoDB, taking them from initial concept to production release.
• Built RESTful APIs and web services, and designed MySQL and MongoDB database architectures to keep the systems performant.
• Gathered requirements directly from business stakeholders and clients and translated them into technical solutions.
• Set up Git branching strategies and version-control workflows, and enforced clean-code standards through code reviews.
• Provided technical analysis, feasibility assessments and architectural guidance for new features and platform enhancements.