Nicola – Typescript, React, AI agent development

A financial services platform built for a growing brokerage business, focused on supporting trading operations, client account management, transaction processing, and internal business workflows. The project involved developing and maintaining scalable backend services, integrating with financial systems and third-party APIs, and delivering features to support the company's expanding operations and customer base.

• Provisioned and operated a bare-metal Kubernetes platform on Talos for a financial brokerage and algorithmic trading system, delivering 50 ms p99 API latency;
• Engineered the infrastructure and execution environment for latency arbitrage algorithms, including server provisioning, colocation, and monitoring, achieving 10 ms p95 tick-to-trade latency (non-HFT);
• Designed and built an automated agentic SDLC pipeline covering code generation, code review, CI/CD, production promotion, automated rollbacks, hotfix deployment, and risk-based approval gates.

An environmental compliance and emissions monitoring platform for industrial businesses, designed to help organizations monitor atmospheric emissions, detect regulatory threshold breaches in real time, automate compliance reporting, and maintain long-term audit records in accordance with EU environmental regulations.

• Partnered with a non-technical founding team to translate regulatory and business requirements into the architecture of a regulated emissions monitoring platform;
• Designed and implemented an on-premises Data Acquisition and Handling System (DAHS) compliant with UNI EN 17255, delivering a solution contracted for deployment across seven industrial facilities.

A digital healthcare platform supporting a network of physiotherapy clinics across Europe, designed to streamline patient care, clinical workflows, and treatment management. The platform supports innovative rehabilitation methods and integrates with proprietary medical devices, enabling practitioners to deliver personalized treatment plans while managing patient records, appointments, and therapy outcomes.

• Defined the architecture, infrastructure, and technology strategy for a medtech company operating without an internal engineering team;
• Developed and launched the company's core physiotherapy platform, now serving 1K+ professionals across 250K+ treatment sessions.

A next-generation SaaS time tracking and workforce productivity platform built to replace a legacy product serving more than 1 million users and a multi-million-dollar annual recurring revenue (ARR) business. The project focused on modernizing the product architecture, improving scalability and user experience, and supporting a seamless migration for an established global customer base.

• Owned the backend architecture and implementation end-to-end, from the initial system design through production rollout;
• Contributed across the full technology stack, including the web application and cloud infrastructure;
• Led the product from proof of concept to a production-ready solution adopted by end users;
• Designed and implemented the platform's AI capabilities, establishing the foundation for AI-powered features;
• Built a natural language chatbot interface, enabling users to interact with the platform through conversational AI;
• Implemented image-based data import capabilities, streamlining user onboarding and reducing manual data entry;
• Established reusable AI development patterns, evaluation workflows, and best practices that became the team's standard for building and validating new AI features.

An enterprise AI knowledge platform that enabled users to retrieve insights from structured and unstructured business data through natural language interactions. Built before Retrieval-Augmented Generation (RAG) became a mainstream approach, the system combined semantic search, knowledge retrieval, and agentic reasoning to help users discover relevant information across internal data sources and business systems.

• Led and mentored two junior backend engineers while supporting the hiring process for backend roles;
• Contributed to the architecture and development of the platform's second and third major product iterations;
• Refactored backend services to support logical multi-tenancy, reducing Kubernetes cluster utilization from 4,000 to 80 pods and lowering compute costs by 60%;
• Scaled the platform's permission management service to process 1,000+ requests per second by leveraging streams and asynchronous generators, reducing both response times and infrastructure costs by 50%;
• Championed the adoption of a monorepo architecture, aligning engineering stakeholders and consolidating 35 microservices into a single repository with a standardized development workflow;
• Drove the adoption of application monitoring and observability practices, enabling teams to identify performance bottlenecks and reduce API response times from 17 seconds to 2 seconds.