Ron Jailall

Providing high-level technical leadership and hands-on engineering for diverse clients, focusing on scalable AI infrastructure and Generative AI product development.

• Custom Vision Language Model & Infrastructure (Matte Model Project): Architected a CPU-efficient, human portrait matting system (MODNet-style/MobileNetV2) using TensorFlow 2/Keras, replacing legacy SDKs in an Electron + React application.
• Designed end-to-end MLOps pipelines on GCP Vertex AI for custom training jobs, dataset management (P3M-10k/Adobe), and seamless ONNX export.
• Implemented CPU-focused inference strategies utilizing ONNX Runtime, enabling real-time alpha matte generation and video compositing with minimal latency.
• Generative AI & RAG Architecture: Designed and implemented real-time RAG-based inference pipelines to power dynamic, personalized product recommendation engines.
• Migrated Nvidia Riva/Triton microservices to AWS, re-architecting the deployment for enhanced scalability and high availability.
• Performance Optimization & Edge AI: Optimized Computer Vision models for embedded Nvidia Jetson platforms using TensorRT and ONNX, achieving significant reductions in inference latency.
• Engineered a cloud-based Stable Diffusion video pipeline for real-time webcam re-rendering and accelerated thumbnail generation.
• Strategic Leadership: Advise client executive teams on AI integration strategies, identifying opportunities to leverage LLMs and diffusion models for competitive advantage.

Led the R&D engineering team in evaluating and productizing cutting-edge Generative AI tools.

• Team Leadership & Strategy: Collaborated with PhD researchers to translate academic findings into production-ready features. Led weekly cross-company technical forums to align engineering and product teams on the latest AI developments.
• Infrastructure Design: Built robust CI/CD pipelines using Terraform, GitLab, and AWS to deploy ML inference servers, supporting microservices architecture across Docker and Kubernetes.
• Model Optimization: Modified C/C++ codebases (llama.cpp & Stable Diffusion Turbo) to suit specific business use cases, optimizing performance for Python-based API endpoints (FastAPI/Uvicorn).
• Product Innovation: Prototyped and deployed React-based demos powered by LLMs and diffusion models, directly influencing the product roadmap and UX design.

Progression from Engineer to Lead, managing critical data pipelines and spearheading the transition to cloud-native technologies.

• Scale & Reliability: Architected and supported data pipelines serving the company's 5 largest enterprise customers (100k+ end users).
• Cloud Transformation: Designed and built an AWS cloud-native Archive utility (Lambda, Batch, Terraform) that was adopted as a core product, generating millions in revenue.
• AI Leadership: Founded and led an internal AI/ML reading group (growing to ~25% of the company) and initiated company-wide AI hackathons to foster innovation.
• LLM Implementation: Designed the inference architecture for a Zoom/conferencing plugin and developed pipelines to deploy GGML Llama models to AWS Serverless architecture.

• Led AV/Room Design Engineering for 200+ learning spaces.
• Developed novel embedded control systems and accessibility UIs; awarded the OIT Award for Excellence.