Dhruvil A. Satani

About Dhruvil A. Satani

I am a Machine Learning Engineer and Research Associate with over 3 years of experience architecting, fine-tuning, and deploying intelligent systems at scale. Technically fluent in Python and core ML libraries (NumPy, Pandas, PyTorch, TensorFlow), I build robust pipelines and prompt-engineered scenarios to power advanced AI applications. At Smodin, I lead end-to-end ML efforts—designing scalable data pipelines, optimizing prompt workflows, and deploying production-ready NLP systems that generate text, summarize content, and deliver insights for thousands of users in agile development environments. At Caltech (Thomson Lab), I’m immersed in innovation—fine-tuning large LLMs for specialized use cases using LoRA and QLoRA, performing embedding analysis, spearheading data acquisition pipelines, and contributing to the FIP (Functionally Invariant Paths) framework for adaptable neural networks. Research• Engineering Flexible Machine Learning Systems by Traversing Functionally‑Invariant Paths – Co-author (DOI: 22•••••34)Core competencies:Python · Deep Learning · NLP · LLM Fine-Tuning · Prompt Engineering · LoRA/QLoRA · Embedding Analysis · Agile Development Future-Facing & Emerging Areas• AI Agents & Autonomous Systems – Building smart, autonomous agents capable of multi-step decision-making (e.g, legal chatbots, AI assistants, production-grade AI agents)• Responsible & Sustainable AI – Integrating ethical guardrails, energy-efficient models, and green computing practices into system design• Collaborative & Memory-Aware Agents – Developing agents with long-term memory, personalized behaviour\'s, and multi-agent ecosystems for complex workflows• No-Code / Low-Code & Vibe Coding – Adopting modern developer methods like “vibe coding” to streamline software production with LLM-assisted prompting workflows Why I’m Unique• Bridging research and product – I translate cutting-edge lab research into scalable, real-world AI systems.• End-to-end delivery – From data collection → algorithm design → model fine-tuning → deployment → MLOps.• Future-ready focus – Adept in today’s tools, prepared for tomorrow’s technologies: AI agents, edge intelligence, and sustainable AI frameworks.

Experience

1. Machine Learning Engineer

   Smodin

   Dec 2021 — Present

   Strategically deployed advanced machine learning models and built scalable, efficient data pipelines to enhance system performance and reliability• Designed and implemented end-to-end ML pipelines, including prompt engineering and optimization for production AI use cases• Managed and maintained code repositories with a focus on clean architecture, version control, and collaborative development practices• Contributed to agile development cycles, driving iterative improvements and cross-functional coordination for timely delivery

Education

• Uka Tarsadia University

  Bachelor of Technology - BTech

  2018 — 2022