Leonardo Lara joined Strat2gyAI through VanHack as an AI engineer. Within six months, his work had already expanded across product development, private AI infrastructure, system architecture, and technical leadership.
Strat2gyAI is an AI-powered strategic growth management platform built to help organizations create, track, and execute strategic initiatives with stronger alignment and clearer decision-making. For a product built around sensitive business context, one of the clearest signs of progress was speed: answers that once took close to a minute could now arrive in seconds.
“The old AI system was taking almost one minute to answer simple questions. Right now, it is almost two seconds. For complex questions, I could reduce it three or four times.”
That improvement was not the result of a surface-level AI feature. Leonardo’s story shows what deeper AI engineering can look like in practice: rebuilding an outdated system from the ground up, creating private large language model infrastructure, improving chatbot performance, and growing into a leadership role as the platform matured.
For a company working with strategy, market, financial, and organizational data, the technical challenge was not simply to add AI. The challenge was to build systems that could understand complex business context, protect client information, and respond fast enough to support executive decisions.
Building AI products for strategic decisions
Leonardo describes Strat2gyAI as a company focused on intelligent strategy actions and roadmaps for organizations, including work around corporate directives, workforce planning, compensation, and strategic vision.
His first major product was an AI chatbot designed to work with client data such as roadmaps, production targets, workforce information, operations data, and strategy documents. From there, his work expanded into a broader portfolio of intelligence tools.
Those products included competitive intelligence using earnings calls, financial data, and news; shareholder activism tracking; market intelligence for emerging technologies and regulatory changes; and executive compensation analysis.
The common thread is clear: help leaders turn complex business information into something they can understand, compare, and act on.
Rebuilding the AI foundation from scratch
One of Leonardo’s first challenges was foundational. The previous AI system was outdated, and the team needed a stronger base for the products Strat2gyAI was building.
That meant changing the AI backend, implementing asynchronous architecture, using tools such as LangChain and LangGraph, building evaluation pipelines for chatbot and retrieval-augmented generation work, and using vector databases to manage context.
The context problem was especially important. Client databases were large and difficult for an AI system to interpret directly. Leonardo had to understand the financial and market domains, then turn messy business data into context the model could actually use.
He also built scraping systems to access external documents and market information that were not available through easy APIs, including sources such as management proxy materials and annual forms.
More context, stronger quality, and client momentum
Speed was only one part of the improvement. Leonardo also says the chatbot now works with 10 times more context, giving the system more information to draw from when answering questions. In strategy work, that matters because useful AI depends on understanding the full business picture, not just returning a fast response.
The improved chatbot and intelligence tools were presented to executives and received strong feedback. As the product matured, Strat2gyAI also secured a contract with a large mining-sector client while Leonardo continued building version two of the AI system with better evaluation metrics.
“We have 10 times more context available to the AI chatbot. For speed, we improved the answer time by more than four times.”
Private AI infrastructure for sensitive client data
Privacy created another layer of complexity. Strat2gyAI’s clients could not send sensitive data through external closed-source AI providers, so the platform needed private large language model infrastructure.
That pushed the work beyond prompt design or API integration. Leonardo had to work through GPU availability, scalability, concurrency, service parameters, coding stability, and infrastructure design.
This is where AI feature development becomes AI systems engineering. The product experience depends on the model, but it also depends on everything around the model: data pipelines, retrieval quality, infrastructure, evaluation, latency, and operational reliability.
What AI engineers need in a startup environment
Leonardo’s advice to other AI engineers is rooted in that same experience. In a startup environment, he says, technical professionals need breadth. Building agents, tools, and retrieval-augmented generation systems is only part of the work.
The role can also demand backend development, cloud services, LLM operations, DevOps, automation, continuous integration, continuous deployment, and infrastructure knowledge.
He also emphasized system design. AI can help write code, but engineers still need to choose the right databases, frameworks, architecture, and infrastructure for the business case in front of them.
That perspective fits the work he did at Strat2gyAI. His value was not only in building AI features. It was in understanding the business problem deeply enough to design systems that could support it.
From contractor to AI engineering lead
The impact also changed Leonardo’s own role. He started as an hourly contractor, working as needed on the project for 20 to 30 hours per week. After six months with the company, he had been promoted into an AI engineering lead role.
“I started in the company as a contractor, working 20 to 30 hours per week. Right now, I got a promotion for an AI engineering lead role.”
That new role moves him further into architecture, planning, timelines, technical direction, and hiring for AI and backend positions.
For companies hiring through VanHack, that is an important part of the story. The right hire does not only complete assigned tasks. The right hire can grow with the business, take ownership of bigger systems, and help shape what comes next.
Why global AI talent matters
AI products need people who can work across the whole system: data, models, infrastructure, product requirements, privacy constraints, evaluation, and business context. Those people are not always sitting in the same city as the company that needs them.
With VanHack, companies can look beyond local hiring markets and find senior technical talent ready to contribute to real business outcomes.
In Leonardo’s case, that contribution showed up in faster AI responses, more context, private infrastructure, new intelligence products, client momentum, and a move into AI engineering leadership.
That is the heart of Built By VanHackers: the value does not stop at the hire. It shows up in the systems, products, and teams that talented people help build after they join.
