Mario Yuwono

Independently designed, built, and launched a visual SaaS platform that enables non-coders to create, customize, and deploy algorithmic trading strategies. Architected a full-stack platform (React/Next.js & Python) translating visual JSON graphs into MQL source code and compiling via MetaEditor; utilized Supabase and Redis for seamless task notifications. Prototyping generative AI features using Google Gemini to translate natural language prompts into structured graphical workflow nodes.

Expanded Docparser product with productionized LLM integrations utilizing the OpenAI API, leveraging prompt engineering to transform unstructured resume data into reliable, structured JSON schemas. Upgraded billing capabilities to Stripe Elements ecosystem, with zero payment failures. Accelerated user onboarding by integrating Amazon Textract into a decoupled SQS/SNS pipeline; designed and implemented a new preprocessing queue to automate rule generation without disrupting existing workflows. Integrated Postmark for high-reliability transactional email routing. Expedited rule parsing configurations by automated table and checkbox detection via OpenCV. Authored a technical research proposal evaluating LLM alternatives (LLaMA, LayoutLM, DeepSeek).

Launched a stable beta product within a tight two-week deadline to secure early client adoption. Introduced project scoping and estimation, improving stakeholder confidence, managing technical debt, and ensuring consistent product delivery timelines. Scaled the engineering team, doubling capacity within two months while mentoring new hires to eliminate development bottlenecks and accelerate platform features.