Ruben Lekli

Turning non-technical client requirements, messy product ideas, and real engineering constraints into production systems. Scope and priority decisions happen early, and stay firm under pressure: what ships first, what gets cut, and what risks breaking if the timeline shrinks. Most projects get built with clients, founders, or team leads who need AI-powered applications, workflow infrastructure, data pipelines, internal tools, new products, or optimizations to existing systems.

Quant dev is part of the same engineering work, with more finance, mathematics, statistics, and optimization in the loop. In practice, that's building research, backtesting, portfolio, and robustness tooling to test market ideas properly, for client work and for strategies traded independently.

AI tools are part of how the work gets done; mostly used for speed and to sketch out scaffolding. Output correctness remains under careful review, ensured by the test harness, the criteria, and a strong oracle, to spot any errors that may slip through. The work still has to be structured clearly, reviewed properly, and maintained like any other production system.