Liquid

Liquid is a financial prognoses platform used by businesses to gain real-time insight into their P&L, balance sheet, and cashflow. The platform consolidates financial data from multiple sources and presents it in a clear, actionable format — helping finance teams make better decisions faster.

My role was Front and Backend Developer, working across the full stack on both product features and infrastructure. The backend runs on Node.js with Express, with MongoDB as the primary database, deployed on AWS using Docker and Kubernetes. On the frontend I worked with Angular, collaborating closely with the UI/UX designers to implement and refine the interfaces.

One of the more interesting features I built was a custom report-builder editor. It allows users to design their own financial reports by combining the same elements the platform uses internally — with realtime data regeneration whenever the underlying data changes. This gave finance teams full flexibility without needing developer involvement every time a report format changed.

I also integrated Stripe for payment processing and led the overhaul of the platform's licensing and subscription structure. Alongside that, I worked on a Multi-Entity Consolidation feature — taking what was a single-entity application and extending it to fully support multiple legal entities, including inter-company transaction elimination and consolidated reporting.

To further support user workflows, I built an AI-powered chatbot using LangChain that can answer financial questions about their prognoses and actually build them in-app too. The user would get a similar experience as Excel, but then with all the power of AI and the platform's data model behind it. This was a very interesting challenge, as it required a deep understanding how LLMs work and how they actually process the data such that it knows what the user wants. It required building a set of skills and tools from which the multiple stages of the AI pipeline could be orchestrated, and then integrating that into the platform in a way that was seamless for the user. The edits were done in real-time over SSE protocol, and the user could see the changes reflected immediately in their reports.

Working at Liquid deepened my understanding of financial domain knowledge — debits, credits, P&L, balance sheets and cashflow — well beyond what typical web development requires. On the infrastructure side I got extensive hands-on experience running production workloads on AWS with Docker and Kubernetes, managing the full deployment pipeline. I grew significantly with Angular on the frontend and MongoDB on the backend, and built real experience with LangChain and AI integration in a production context. Building the report-builder also taught me a lot about designing flexible, user-configurable systems without sacrificing reliability.