Nandni Jaiswal
Hi. I’m a final-year Computer Science student at IIT Patna and an independent ML researcher. I spend most of my time on one question: can thinking differently make machines reason better? Most recently I was a strategy and operations engineer at ArisX, building agentic systems for financial institutions.
What I’m working on
I’m interested in whether alternative cognitive strategies give a computational edge in machine reasoning: novelty seeking, spotting traps, exploring instead of exploiting, generating hypotheses. I turn those hunches into reproducible systems and benchmarks, so we can actually tell whether they work.
Broadly: machine reasoning, cognitive architectures, agentic AI, scientific discovery, and how humans and AI make decisions together.
Things I’ve built
NeuroArxiv · STEM reasoning & research engine
A reasoning framework extended into STEM problem solving and scientific literature exploration. It pulls from arXiv inside the reasoning loop, so the system reads what already exists before proposing a new architecture or solution. 400+ stars, 10,000+ downloads.
Runtime Compliance Infrastructure · auditable AI in healthcare
Built at the IIT Patna Incubation Centre. Healthcare AI workflows make decisions at runtime; RCI keeps an auditable compliance trail of them, so what the system did stays inspectable after the fact instead of being a black box.
Agentic research & profiling pipeline · built at ArisX
An LLM-powered pipeline for automated research and customer profiling. Extraction, synthesis, and workflow orchestration together took 30+ hours a week of manual work down to under 1.25 — about a 96% cut.
Where I’ve been
ArisX, strategy and operations engineer. Bangalore. VoiceAI, NBFC workflows, and agentic automation — AI systems for high-volume, decision-intensive work inside financial institutions. Mostly turning fragmented information into something business and sales teams can act on.
Incubation Centre, IIT Patna, AI research intern. Runtime compliance infrastructure so healthcare AI workflows leave an auditable trail.
Exthalpy, co-founder and researcher. Bangalore. Tried to represent how experts actually make decisions and run consultations, computationally — building on earlier work with healthcare professionals. $25,000 in grants and incubation from MeitY, Government of India, through DRISHTI CPS, IIT Indore.
Independent research. Whether alternative cognitive strategies help machines reason — and turning those hypotheses into benchmarks, systems, and open-source research infrastructure people can actually run.
Misc
- Research funded by the Ministry of Electronics & Information Technology, Government of India.
- Python, system design, CI/CD, and open-source development; experimental design, benchmarking, literature review, and cognitive modeling on the research side.
- B.S. Computer Science and Data Analytics, IIT Patna, 2023–2027. CGPA 8.90.