About Me
I am an AI Research Engineer passionate about Physical AI and dexterous robotics. As a Data Science graduate from Hanyang University, my work sits at the frontier of Robot Foundation Models and Vision-Language-Action systems; spanning 3D scene-aware perception, tactile-torque sensory integration, and touch-aware dexterous manipulation policies. Beyond robotics, I bring strong hands-on experience in machine learning, NLP, and statistical modeling, having built and deployed ML systems across AI robotics labs and IT startups; from patent-filed anomaly detection to sentiment analysis at scale. I am driven to push the boundaries of what intelligent systems can do in the physical world.
Technical Skills
Languages
Python, C, C++, R, JS, TS
Frameworks
TensorFlow, PyTorch, scikit-learn, Keras, NumPy, pandas, Django
Robotics
Isaac Sim/Lab, Robot Manipulation, 3D Object Scanning
Dev/Analytics
Angular 10, ELK Stack, PsychoPy, API Development
Education
Bachelors of Engineering in Data Science
2022 - 2026
Hanyang University, South Korea
- Major: Data Science Engineering and Physiological Brain Science
- GPA: 4.0 out of 4.5
Middle and High School
2015 - 2022
Vibgyor High School, India
- Major: Science (Mathematics, Physics, Chemistry, English, Physical Education)
- 2022 - Ranked in the Top 5 in Grade 12 (92.6%)
- 2020 - Ranked in the Top 10 in Grade 10 (93.5%)
Primary School
2008 - 2015
International School of Koje, South Korea
- Studied in an international environment
- Experience of solving practical case studies
Experience
AI Research Engineer Intern
Dec 2025 - Aug 2026
Data Team
- Built and curated robotic manipulation datasets in Isaac Sim, optimized dataset composition to achieve equivalent model accuracy with 50% reduced data volume.
- Fine-tuned and systematically benchmarked multiple VLA models (GR00T, π0, π0.5, BC-RNN-GMM), establishing performance baselines for robot foundation model training.
- Developed an end-to-end automated training and evaluation pipeline across Slurm and Kubernetes GPU clusters, reducing model setup time from 20 to 30 minutes to under 30 seconds.
Physixel Team
- Built simulated tactile and joint-torque sensing for dexterous bimanual manipulation in MuJoCo, and produced a multimodal robot-learning dataset for training world-action models and VLA policies.
- Integrating 3D geometry perception into VLA models via foundation model-based scene understanding, enabling multimodal sensor fusion (tactile, joint torque) for touch-aware dexterous manipulation policies.
Data Science Intern
Mar 2025 - Dec 2025
- Part of the R&D team developing machine learning–driven multi-module system for fake account detection and anomaly detection, patent filed for the technology.
- Built and evaluated deep learning image classifiers using Residual Networks (ResNet) for synthetic image detection, targeting AI-generated profile photos and fraudulent user identities.
- Developed and evaluated machine learning models for natural language processing tasks using BERT, with a focus on sentiment analysis and text classification to support improvements in a social networking app (Moji) and provide user engagement insights.
- Experience Letter
Undergraduate Thesis Research
March - Dec. 2025
Advisor: Prof. Noh Youngtae, Department of Data Science
- Research aimed to enable non-invasive continuous health monitoring by addressing noise interference in mmWave FMCW radar (AWR1642BOOST) vital sign signals.
- Developed a CNN+BiLSTM+Attention model for contactless heart rate and breathing rate estimation from mmWave radar signals; 3× improvement over classical signal processing baselines.
Teaching Assistant
Sep. 2024 - Dec. 2024
- Managed and tracked a class of over 30 students for the course “Understanding Business Environment”.
- Helped facilitate group discussions and ensured smooth execution of class activities.
- Experience Letter
Software Engineer Intern
July 2024 - Aug. 2024
Bespoke System Pvt. Ltd., India
- Worked on web development using ASP.NET framework, MS SQL, Angular 10, Web API, and SQL Server.
- Gained experience in both frontend and backend development, focusing on API integration and database management.
- Experience Letter
Projects
1) Contactless Vital-Sign Monitoring: Benchmarking Denoising Frameworks for mmWave Radar
March - Dec. 2025
- Benchmarked three denoising frameworks (BPF+FFT, DWT, CNN+BiLSTM+Attention) for mmWave radar heart rate and breathing rate estimation.
- Deep learning model achieved MAE of 3.12 (heart rate) and 2.99 (breathing rate), significantly outperforming classical signal processing baselines; enabled driver stress detection and clinical health monitoring applications.
2) StressMate – Mental Wellness Web App
May 2025 - June 2025
- Developed StressMate, an AI-powered emotional companion web app designed to support mental wellness through empathetic conversation, emotion tracking, and personalized wellness insights.
- Built a Django-based app integrating LLMs (Perplexity AI, Gemini) along with RAG. Deployed on AWS using EC2 and S3 for scalable hosting and cloud based storage.
3) PlateCheck Application
Oct. 2024 - Dec. 2024
Developed LLM based Korean Food Ingredient Recognition that accommodates both food image and text inputs using Django. Integrated a fine-tuned ResNet model for image-based dish identification and utilized GPT-4o Mini for extracting relevant ingredient details. Evaluated the service based on HCI factors via user study.
4) Food Demand Forecasting using Temporal Convolutional Networks
Oct. 2024 - Dec. 2024
Built a time series forecasting model using Temporal Convolutional Network (TCN) to forecast food demand based on historical sales data.
5) News Audio Summarizer using BART
Sep. 2024 - Dec. 2024
Designed an NLP-based system for converting audio news content to text and generating concise summaries by fine tuning the BART model on transcribed news articles paired with human written summaries.
6) Perceptual Misinformation and Memory
March 2024 - June 2024
Designed and conducted a cognitive psychology experiment investigating how different types of perceptual misinformation influence visual memory, analyzing recall accuracy and reaction times.
Publication & Conferences
RLDX-1 Technical Report
Seoul, May 2026- Contributed to research on Robot Foundation Models and Vision-Language-Action (VLA) systems for dexterous humanoid robot manipulation.
- Supported the development and evaluation of simulation-based manipulation, pick-and-place, and humanoid robot learning tasks.
Presented research poster, “Cerebellum Activation in Depression”
Seoul, Aug. 2025Korean Society for Brain and Neural Sciences (KSBNS) Conference
The 12th Swiss-Korean Life Science Symposium, Research Attendee
Seoul, May 2025Scholarships and Achievements
Excellence Award, College of Engineering, Hanyang University
August 2026Awarded at graduation in recognition of consistently outstanding academic performance throughout four years of undergraduate study
Daewoong Foundation Global Scholarship
Nov. 2025 - March 2026Selected as a top global talent from international students in South Korea based on career potential and growth mindset
Hanyang Brain (Academic Excellence) Scholarship
Awarded 3 consecutive semesters
March 2025 - March 2026Merit-based scholarship awarded to top-ranked students in the department for outstanding academic performance and high GPA
Special Scholarship for Academic Encouragement Hanyang (Foreigner)
March 2024 - June 2024Merit-based scholarship awarded for high academic performance among international students at Hanyang University
Languages
Native
English, Hindi
Intermediate
Gujarati
Beginner
Korean