Arya Maran
Bridging AI and Computer Science
Developing intelligent systems that help us understand and protect our world through the power of machine learning.
AI Projects/Research
Always looking to further my expertise in the field of AI through projects and research.
Climate LLM
Advanced language model fine-tuned on climate science literature, enabling natural language queries about environmental data and climate predictions.
Bloompoint
Full-stack spatial ML application detecting algal blooms from 20 years of NASA satellite data, with a RAG-driven attribution engine to correlate anomalies with real-world industrial sources.
Interested in collaborating on AI research or projects?
My Journey
About Arya
I’m a CS student at Purdue focused on machine intelligence, driven by a passion for AI that creates measurable real-world impact. From fine-tuning large language models and building spatial ML systems for environmental monitoring, to explainability research on robotic manipulation — I work at the intersection of research and engineering.
- Education
- BS Computer Science (Machine Intelligence track) with a Minor in Mathematics at Purdue University, Class of 2027. Coursework spans Data Structures & Algorithms, Data Engineering, Linear Algebra, Systems Programming, and C.
- Research & Industry
- Undergraduate researcher at Purdue's Data Mine (LPR systems with YOLO/CNNs for V2X Inc.) and the University of Florida (explainable AI for robot manipulation using gradient-based attribution). Previously AI/ML intern at Crcle AI, fine-tuning 7B-class LLMs with QLoRA — cutting peak VRAM by 75%.
- What Drives Me
- I want to build AI that matters — from spatial ML catching algal blooms to language models that make climate science accessible. Currently also consulting on web optimization and UI/UX for SHARE Cancer Support, helping patients navigate a women's cancer support network.
Let’s Connect
Interested in collaborating on AI or just having a chat? Have questions about my research? Let’s start a conversation.
Or reach out directly