Socials

Project Concept

PrivacyLens is an on-device “privacy gate” for creators: it scans a raw phone video offline to detect sensitive visual + audio information (faces, screens, addresses, IDs, phone numbers, emails, spoken PII), then generates timestamped alerts before you upload to any social platform.

Entry

Status: Submitted

Last saved: February 21 at 6:13 PM PST

View Entry

Team Roster

You must be registered for the event to view the team message board.

Jidhnyasa Mahajan Team Lead RSVP Approved

AI/ML Engineer at Independent Researcher & Builder

I led the project from ideation to delivery—defining the PrivacyLens problem statement, shaping the local-first edge–cloud architecture, and planning the build to align with all judging rubrics. I owned the pitch narrative and demo flow, ensured the hybrid escalation story was clear, and created the documentation (README, demo script, and architecture overview) so judges and teammates could understand and run the project instantly.

AI Engineer (Research & Development) who is self-employed and works as a Content Creator for YouTube, based in the San Francisco Bay Area. With 4 years of experience, their background includes education from Stanford University, the University of Georgia - Franklin College of Arts and Sciences, and the University of Mumbai, focusing on fields like AI Professional Program (Deep Generative Models), Computer Science, and Computer Engineering. Their projects involve producing AI-focused YouTube content on generative AI creativity and AI-generated music, alongside technical work in style transfer, differential privacy optimization, and developing full-stack prototypes. They are looking for co-founders, investors, speaking opportunities, and sponsorships, and can help with GTM, partnerships, product review, technical architecture, and developer marketing.

Generative AI creativity, AI-generated music, deep learning, NLP, differential privacy, style transfer, technical architecture, product review, GTM, partnerships, developer marketing, technical co-founding, business co-founding, investment, speaking opportunities, sponsorships

Projects include producing AI-focused YouTube content on generative AI creativity and AI-generated music. Technical implementations feature style transfer using DenseNet and Adaptive Instance Normalization, differential privacy optimization experiments, and R Shiny applications for epidemic modeling. Other work involves developing full-stack prototypes with ReactJS, Spring Boot, and Docker, and executing NLP tasks such as NER and dependency parsing using TensorFlow and Keras.

Akshay Aralikatti RSVP Approved

Student at CSU Chico

Akshay owned the leaderboard track—driving our generate_hybrid() routing to maximize on-device correctness, speed, and tool-call reliability on the hidden benchmark. He iterated aggressively with benchmark.py, tuned confidence thresholds/tool selection, and hardened validation + repair loops to push toward near-perfect scoring. He also focused on fine-tuning and calibration strategies to improve FunctionGemma’s tool-call precision.

Interested in contributing to impactful tech

Agentic AI, Robotics, Autonomous Systems

https://akshayarali.github.io/akshayaralikatti/ akshayarali.github.io/voice_bot/

Srushti Jagtap RSVP Approved

Software developer at Locomex

Srushti built the core agentic workflow by orchestrating the right local tools end-to-end and ensuring seamless execution on-device. She integrated video/audio pipelines using OpenCV for frame sampling + redaction rendering and Whisper (speech-to-text) via cactus_transcribe for low-latency voice commands. She also handled tool schemas, function-call execution loops, and made the pipeline robust for real demo inputs.

I am a computer science graduate student at NYU Tandon with a strong focus on full stack and AI driven systems. I build products that mix clean user experiences with solid backend design. My work centers on AI agents, voice interfaces, cloud systems, and real time data pipelines. I have built tools for legal intake, ESG reporting, workflow automation, and customer support using LLMs, AWS, Kubernetes, and modern web frameworks. I enjoy turning complex problems into practical software that is fast, secure, and easy to use.

python , aws, frontend , vibe code

• Personal Injury Agent I am building an AI legal intake system with agent based workflows. It uses LangGraph for multi step reasoning, a visual React Flow UI, and live voice input. The backend runs on Kubernetes and scales with traffic. • AI driven RFP and ESG automation at LOCOMeX I am improving systems that scrape external data, run ETL pipelines, and use LLMs to speed up RFP decisions and ESG reporting. The focus is accuracy, speed, and clean reports.