Team
UX Machina
Project Concept
No description has been added yet.
Entry
Status: Submitted
Last saved: February 21 at 5:44 PM EST
Team Roster
You must be registered for the event to view the team message board.
Manoj Sadanala Team Lead RSVP Approved
CS Grad Student at Stevens Institute of Technology
Manoj designed and implemented the adaptive UI system using React (Vite + TypeScript) and CopilotKit to bridge structured AI outputs into dynamic interface changes.
He shattered traditional chat-based UX conventions by building:
A morphing UI that changes layout density based on emotional signals
A reducer-driven adaptive state engine for deterministic transitions
Voice-driven interaction using the Web Speech API
Emotion-responsive theming (color, motion, layout compression)
Structured UI command parsing layer
Manoj ensured that Claude’s reasoning manifests visually — transforming intelligence into interface. Rather than displaying AI text in a chat window, the frontend renders evolving mission controls, adaptive overlays, and dynamically generated interaction patterns.
As team lead, he coordinated cross-component integration to ensure Tavus emotion signals, Claude directives, and backend state transitions worked seamlessly together.
I’m an M.S. Computer Science student at Stevens Institute of Technology with a background in Electronics & Communication Engineering. I focus on building practical, real-world systems across applied ML, embedded sensing, and full-stack development. My work spans wearable sensor platforms for human movement analysis, mobile and web applications, and data-driven tools for operations and healthcare contexts. I enjoy fast-paced environments, collaborating with technical teams, and shipping ideas from concept to deployment.
Applied AI in healthcare, wearable sensing, human-centered systems, and data-driven decision making. Interested in collaborating on ML-enabled products, embedded + software hybrid systems, early-stage startups, and research-to-product transitions. Open to connecting with builders working on impactful, real-world problems.
I’m building end-to-end applied systems that bridge software, ML, and hardware. Current projects include the design and integration of embedded wearable sensing systems for gait monitoring, where I lead hardware development, sensor integration, firmware design, and real-time data acquisition, and a modern Next.js e-commerce platform. I enjoy rapid prototyping, building data pipelines, and turning research ideas into working products.
Gladys Preysler RSVP Approved
founder at epsilondelta
Gladys architected the intelligence layer using Claude (Anthropic API) as a structured reasoning engine rather than a conversational chatbot. She designed prompt scaffolds and JSON schema enforcement so Claude generates UI directives, phase transitions, option sets, and tone adjustments instead of free-form text.
She also integrated Braid to manage structured reasoning flows and deterministic output parsing.
Her contributions included:
Designing Claude system prompts for phase-based mission logic
Enforcing JSON output schemas for stable UI rendering
Building the adaptive mission engine (Infiltrate → Breach → Escape)
Implementing complexity scaling based on emotional inputs
Creating the “cognitive load balancing” logic
Gladys transformed Claude into a generative experience director that dynamically composes the interface.
Software engineer with experience building enterprise applications across the stack
Leverage AI for fintech
An AI-based educational tool
Jithendra Puppala RSVP Approved
Student | Data Scientist at New York University
Jithendra built the orchestration backend using Node.js, WebSockets, and a structured state engine to synchronize Claude outputs, Tavus emotion signals, and frontend updates in real time.
His responsibilities included:
Designing the WebSocket gateway for low-latency UI synchronization
Implementing session-based state management
Building the mission timer and deterministic phase transitions
Creating the API bridge between Claude reasoning outputs and frontend UI commands
Implementing demo-mode fallback architecture
He ensured system stability by separating AI reasoning from rendering logic, preventing hallucination-driven UI instability.
A Data Scientist with 2+ years of industry experience at Jio Platforms, developing and deploying ML models that improved precision/recall by 25%+ and scaled to 150K+ IoT devices. Currently pursuing an M.S. in Computer Science at NYU, focusing on Machine Learning, Computer Vision, and Data Science. Seeking a Summer 2026 Data Science/AI/ML internship to apply expertise in modeling, deployment, and scalable analytics.
Deep Learning, Computer Vision
Anh Vu RSVP Approved
Software Engineer at VeneraAI
Anh led the real-time emotional perception pipeline using Tavus APIs for video-based facial signal interpretation. She integrated Tavus’ emotion and facial-expression inference layer into our WebSocket orchestration backend, allowing live webcam feeds to influence UI density, tone modulation, and mission pacing.
Her responsibilities included:
Implementing Tavus video stream ingestion and response parsing
Mapping emotion confidence scores → structured UI state modifiers
Designing the “stress compression” logic that dynamically reduces interface complexity
Creating fallback mechanisms when inference confidence drops
Calibrating emotional thresholds to avoid jitter and ensure stable adaptation
Anh’s work made the interface psychologically adaptive instead of purely reactive.
I’m currently a full-stack software developer at a healthcare startup, building a mobile app for personalized healthcare management.