About Me

KaleidoForge is a studio created by me, Romina Molina, to connect machine learning, optimization, and real hardware systems. My background spans applied research, FPGA acceleration, embedded intelligence, and full ML-to-hardware workflows.

My work sits at the intersection of machine learning, optimization, and hardware acceleration. Over the years I’ve worked across the full ML-to-hardware stack: training, compression, FPGA deployment, profiling, and workflow design.

What I enjoy most is making these systems clear, structured, and practical, so that teams can reason about them instead of fighting them.

I help researchers, engineers, and organizations connect ML ideas with hardware reality, building workflows that are reproducible, explainable, and aligned with constraints.

Why KaleidoForge

KaleidoForge emerged from a simple idea: modern ML systems don’t need to be complicated to be powerful. When tools, workflows, and hardware interact coherently, you get systems that are easier to maintain, faster to iterate, and more intuitive to extend.

It’s a space focused on clarity, engineering discipline, and thoughtful design, without unnecessary friction.

Why Work With Me

• Experience across the ML workflow: training, optimization, compression, FPGA acceleration.
• Ability to make advanced concepts concrete and understandable.
• Strong focus on documentation, reproducibility, and reasoning.
• Guidance adapted to real constraints and real projects.
• Technical depth without unnecessary complexity.

Vision

To help people build ML systems that run efficiently, scale sensibly, and remain understandable over time. Clarity is the foundation for reliable engineering.

Values

• Clarity: systems and ideas should be explainable.
• Iteration: progress through small, meaningful steps.
• Coherence: decisions aligned with constraints and purpose.
• Access: knowledge should be practical and shareable.

Design Principles

Every project, course, and workflow within KaleidoForge follows a set of principles that keep technology purposeful and human-centered.

• Human-Readable Systems: clarity first, always.
• Hardware-Aware Intelligence: algorithms shaped by real constraints.
• Reproducible by Design: workflows that can be rebuilt and understood.
• Creative Logic: structured engineering guided by curiosity.
• Fast Meaning: efficiency aligned with purpose.