
AI, Robotics and Law | TKS @ NN Scholar | Researcher @ Perrin | ISEF'23 | IRIS 23' 24'
I'm a Research Associate at The Perrin Institute living in 🇮🇳 New Delhi, India. I specialize in ethical AI governance frameworks for emerging economies, with a focus on developing accountability mechanisms for public-sector AI implementations. I'm also passionate about aerospace research, environmental monitoring, and innovative technology solutions.
I'm currently building a first-of-its-kind AI audit framework — a dynamic system that continuously monitors bias, transparency, and societal impact of deployed AI models. I'm also an innovative student at TKS for the class of 2025, where I focus on Artificial Intelligence (AI) and Brain-Computer Interfaces (BCI). If you'd like to collaborate on research or projects, you can connect with me on GitHub.
You can read my blog or follow me on these platforms.
My journey from a small Indian town to global innovation stages is a story of resilience and creativity. Raised with limited resources, I turned constraints into opportunities—spending hours in a modest school lab, driven by curiosity and purpose.
My perseverance led to national innovation awards, aerospace research publications, and recognition at Regeneron ISEF. With a focus on real-world impact, I built drones for ecological monitoring and AI solutions for societal problems. A full scholarship to The Knowledge Society (TKS) marked a turning point, where I thrived among global innovators.
At TEDxBistupur 2025, I spoke on "From Constraints to Creativity," showing how resourcefulness can spark powerful innovation. Watch my talk to learn how I transformed limitations into opportunities for growth and impact.





_Dhanbad_Logo.svg.png)

.jpeg)
.png)






.png)
















.png)
.jpeg)
.jpeg)
Selected as 1 of 30 Research Associates globally (out of 3,500+ applicants) to produce original research influencing federal agencies, foreign ministries, and international forums like the UN General Assembly. View my expert profile
Delivered a powerful talk on "From Constraints to Creativity" at TEDxBistupur, sharing my journey of transforming limited resources into innovative solutions that address global challenges in AI sustainability and environmental technology.
Architected a "Triple-Stack Optimization" framework, integrating frequency-aware GPU scheduling, algorithmic energy minimization, and lightweight AI models to address AI's energy crisis in Azure data centers; leveraging the LATEST tool for dynamic GPU scaling, Dual-Partition AVF methods for HPC efficiency, and 16-bit quantization with 2:4 pruning for leaner LLMs, projecting $3.8B in annual savings and a 2.1M ton COâ‚‚ reduction.
Formulated the "Nanofluidic-Adaptive Bipolar Osmotic-Responsive (NABOR)" framework, proposing the use of engineered MoSâ‚‚ membranes for enhanced water flux and adaptive energy scheduling for optimal renewable integration.
Completed an 8-month internship at the Indian School of Business (ISB), Mohali, during Grade 11, gaining valuable insights into my national-winning project hosted by Atal Innovation Mission (AIM), NITI Aayog, Government of India.
Mentored by the amazing Rhys Lindmark, I overcame procrastination, improved focus, and achieved a disciplined work ethic of 15+ hours/day. Currently writing a book with 80 pages completed so far, dedicating 2+ hours/day to the project. Awarded a Full-Ride Scholarship ($6,000 USD) to attend the Roote Fellowship.
My research spans AI governance, aerospace technology, environmental monitoring, and cognitive enhancement, with publications in peer-reviewed journals and presentations at international conferences.
Pandey, S. (2024). Accelron Aerospace Journal, 3(3), pp. 487–494. doi: 10.61359/11.2106-2446
View paper →Pandey, S. (2024). Accelron Aerospace Journal, 2(5), pp. 287–294. doi: 10.61359/11.2106-2421
View paper →Pandey, S. (2025). The Perrin Research Institution Working Papers / SSRN, 37 pages. doi: 10.2139/ssrn.5433197
Pandey, S. (2025). SSRN Working Paper Series, 29 pp. doi: 10.2139/ssrn.5399515
Pandey, S. (2023). Regeneron International Science and Engineering Fair (ISEF), Behavioral and Social Sciences Category, Booth ID: BEHA014.
View abstract →Pandey, S. (2025). Pew Research Center, 15 pp. (Under Review)

My passion for innovation was ignited at The Knowledge Society (TKS), a global community that operates like a talent investor for young minds. As a NN Scholar, I immersed myself in the world of Artificial Intelligence and Brain-Computer Interfaces, building over five AI platforms and collaborating with peers who share a vision for positive change.
I've published research papers and built innovative projects that address real-world challenges in environmental monitoring, space technology, and public health.
AeroShield - An advanced air purification system I developed that's specifically engineered to destroy airborne pathogens and pollutants. Unlike traditional filters that merely trap particles, AeroShield employs a sophisticated filtering mechanism composed of a porous 3D network of graphene-coated zinc oxide (ZnO) nanostructures, capturing over 95% of particles larger than 1 micron. The system uses Joule heating to rapidly raise the filter's temperature to approximately 300°C, decomposing harmful microbes into harmless substances.
UAV-based Relaying and Distributed Machine Learning - My research paper introduces a novel split learning framework that optimizes IoT inference workloads across ecological monitoring networks. The study demonstrates 30% faster convergence and 47% reduced energy consumption compared to traditional approaches. View paper
DivyaSat Platform - This groundbreaking paper presents a low-cost, high-efficiency nanosatellite designed for environmental monitoring. Using quantum-resistant cryptography and edge-based computer vision, DivyaSat analyzes deforestation patterns and biodiversity metrics in real-time with 93% accuracy at 1/8th the power consumption of conventional satellites. View paper
Face Recognition Attendance System - In 9th grade, I created this system using Python and the KNN Algorithm, which was recognized by the Atal Innovation Mission. This project automatically marks attendance using facial recognition technology. View on GitHub
Here are some of my other notable projects:
From the spark of an idea to real-world impact, my journey as a builder and innovator has been filled with curiosity, creativity, and a love for solving meaningful problems.
It all began with Sheero AI, India's first multimodal AI voice agent. I dreamed of a mentor that could do it all—analyze screens in real time, interpret video feeds, and even tutor students on the web. Sheero became a true companion: drawing math graphs on the fly, critiquing LinkedIn profiles, detecting book titles from a video, and offering support as both a therapist and a tutor. Watch the demo.
Inspired by the wonders of biology, I set out to reimagine data itself. The Genomic Hyperstructure Framework was born—an architecture that encodes information like DNA, with synthetic nucleotide pairs and a double-helix for bidirectional access and self-replication. It's a leap toward the future of healthcare, secure storage, and AI. See it in action.
During a hackathon, I built PollyMath, an AI-powered anti-cheating exam platform. Imagine turning any PDF into an auto-graded test, with live proctoring and anti-cheating features that keep everyone honest. Over 500 students have already used it to prep for exams! Try the live demo.
Learning should be fun and interactive, so I created VoiceBook. It transforms static materials—PDFs, YouTube videos, documents—into lively, podcast-style audio. Two AI personas, Alex and Sarah, bring the content to life, answering questions and making learning a conversation. Listen to a demo.
Curiosity led me to build the Open Research Canvas, a context-aware AI chatbot that can analyze anything: videos, images, podcasts, documents. With a drag-and-drop canvas, users create knowledge graphs and get hyper-specific insights, whether it's summarizing a TED Talk or analyzing a gene-editing diagram. Watch the demo.
Finally, with VisionSense, I tackled a real-world health challenge. This deep learning tool diagnoses retinal disorders from fundus images with 92% accuracy, helping clinics in resource-limited regions reduce diagnostic delays by 60%. See the results.
Here's some tools, technology and products I use every day.
Warp
Modern, Rust-based terminal with AI.
•Untitled UI
The ultimate UI Kit for Figma.
•Typefully
Professional tools for Twitter.
•Clay
The interpersonal CRM.
•Wealthfront
Grow your wealth with automated investing.
•Robinhood
Commission-free stock trading.
•Superpower
The next-generation digital clinic.
•July
Premium Australian luggage.
•Waymo One
Summon a self-driving car.
•Visual Electric
AI-powered, photo-realistic image generation.
•Mercury
Simple banking for U.S. startups.
•ElevenLabs
AI speech synthesis and voice cloning.
•Cal.com
The better way to schedule your meetings.