About Me.

Meta — New York

Product Strategy Lead

• Optimized Instagram Creator Monetization product; developed performance-based business models.
• Led product research and design sprints on Instagram interactivity; initiated new gesture interactive formats, increasing engagement efficiency by 4%.
• Conducted growth research and content strategy for Facebook VP Analytics on Social Video and Creator.
• Drove a team of 9 in Meta Hackathon; developed Creator Collaboration Portal and won annual winner.

TikTok — New York

Sr. Product Marketing Manager

• Managed Shop Ads growth across 5 internal teams and 3 external channels, capturing a 10× growth rate.
• Developed annual product analysis and led the 2022 roadmap and 4 core experiments programs.
• Led UX research on social commerce; analyzed user behaviors and mental models.

TikTok — New York

Product Marketing Manager

• Initiated go-to-market strategies for branding ads, resulting in MoM growth of 45%.
• Managed pricing strategies for premium reservation ads with annual $730M revenue portfolio.
• Synthesized inbounding product strategy with cross-functionals; led analysis of product research.

Harvard · Wearable Systems · Product Design · Speculative Design

SIXTH: Physical Intelligence Wearable for Female Athletes

• Developed full sensing and signal-processing architecture for SpO₂, HRV, core temperature, and sweat sodium channels.
• Designed on-body actuation layer combining haptic actuators, thermal controls, and audio output.
• Fabricated electronic-textile assemblies via laser cutting, digital embroidery, machine-sewing, heat pressing, and precision wiring.

MIT · Physical AI Hackathon — Winner · Textile + Electrical Engineering

ConnectQ: Embodied AI as a Textile Interface

• Developed a knitted conductive wrist wearable capturing continuous physiological signals: temperature, position, and pressure.
• Contributed to LLM-based affective inference pipeline mapping multimodal physiological and gestural inputs to emotional state classification.
• Integrated Arduino Uno Q with touch sensor, IMU, motor, and haptic actuator subsystems.

Harvard · Advised by Patrick Slade, Harvard Slade Lab · Biomechanics · Wearable Systems

Electrical Muscle Stimulation for Lower Limb Coordination

• Designed within-subject dual-task paradigm across four counterbalanced conditions combining cognitive load and calf TENS stimulation.
• Processed bilateral knee and ankle joint angle time-series from OpenCap; computed three coordination metrics normalized to individual baselines.
• Designed modular wearable attachment system addressing inter-participant variability and personalized EMS.

Harvard · Advised by Prof. Yoel Fink · E-Textile · Edge AI · Physical Computing

Tactiq: Pressure Sensing Textile for Healthcare

• Developed woven textile with high-resolution sensing for clinical deployment using ESP32 and EasyEDA circuit schematics.
• Trained a 4-class posture classification neural network achieving 85.7% real-time accuracy, 0.85 weighted F1 score, and ~1 ms on-device inference latency.
• Evaluated silicone formulations, conductive yarn types, velostat geometries for signal quality and fabrication repeatability.

Harvard · Fabrication · Rehabilitation Engineering

Inflatable TPU Footwear for Rehabilitation

• Designed and fabricated a pneumatically actuated insole using flexible TPU filament with a lattice architecture.
• Developed custom G-code scripts to parametrize print geometry per patient morphology.
• Architected full actuation system: pneumatic channel layout, pressure zone mapping, and inflation sequencing logic for gait rehabilitation.

Harvard · Speculative Design · BioDesign · Affective Computing

Algorithmic Morphogenesis

• Acquired and processed 4-channel EEG data via Muse 2; applied denoising, frequency-band isolation, and spectral analysis to derive two composite indices.
• Mapped physiological time-series onto a 2-axis robotic mechanism encoding cognition as a dynamic spatial and luminous signal.
• Fabricated complete robotic plotter with Arduino Uno integrating neural control, servo actuation, and programmable lighting in response to human emotion.

Harvard · Advised by Prof. Rosalind Picard & Prof. Paul Liang · Affective Computing · Multimodal AI

AI-Generated Affective Scent

• Investigated olfaction as an intervention medium for simulated driving scenarios.
• Proposed system where AI infers emotional and cognitive state in real time and generates scent prescriptions to modulate driver affect.
• Designed multimodal reasoning pipeline with VLM and LLM-inspired scent generation.