Trisham Patil

About Me

My Journey

I started as a mechanical engineer. Long before I wrote production code, I was trained in mathematics, physics, thermodynamics, and fluid mechanics, and in the habit of reasoning about problems from first principles. That background taught me how physical systems actually behave: how they are constrained, how they fail, and how to model them with numerical methods and computation. It also gave me a systems-thinking instinct that still shapes everything I build today. Mechanical engineering was never something I left behind; it is the foundation the rest of my work is built on.

Software did not replace that foundation, it grew out of it. During graduate work in thermofluids and computational fluid dynamics, simulation and scientific computing pushed me from solving equations by hand into writing code, first in MATLAB and then increasingly in Python for numerical simulation and data analysis. What began as computational engineering gradually became broader software engineering: I moved from one-off scripts to repeatable data pipelines, backend services, and systems designed to be reliable and to scale. Software turned out to be the most powerful tool I had for building and scaling engineering systems.

That same trajectory carried me into machine learning and AI. Working with computation and data made the transition natural: from data pipelines and model training into NLP, deep learning, transformers, and LLMs. I became interested not just in using models but in engineering the systems around them, including RAG and retrieval, agentic orchestration, inference serving, GPU-aware optimization, and the backend infrastructure that keeps production AI reliable. I approach AI the way I was trained to approach any engineering problem: as something that must be measurable, constrained, and reproducible.

The more deeply I worked across software and AI, the clearer it became that my mechanical engineering background was not a detour. It was the other half of the problem. AI, software engineering, physics, and mechanical engineering are not separate career paths competing for my attention; they are complementary disciplines. Modeling a physical system, writing the code to simulate it, and training a model to reason about it are the same problem viewed from different angles, and having worked on all of them lets me treat them as one integrated engineering challenge.

My long-term direction sits at the intersection of AI and physical engineering: building intelligent systems that can reason about, simulate, optimize, and control complex physical systems. Concretely, that means work like physics-informed neural networks, scientific machine learning, and digital twins, and applying those methods to domains where physics, computation, and engineering all matter at once, from autonomous systems and robotics to satellites, spacecraft, and rocket propulsion. I want to build AI systems that interact with the physical world and help solve hard engineering problems, combining mechanical engineering, software engineering, and AI into a single way of working.