2026–Present
Muro AI
Founding AI Engineer San Francisco
I build the language model systems, document intelligence, and enterprise integrations behind PrecOS. My work spans training, fine-tuning, real-time inference, model optimization, and evaluating agent tools before release.
I built read and write integrations with Procore, Autodesk BuildingConnected, Microsoft Teams, and Microsoft Dynamics 365. In document pipeline benchmarks, indexing ran 5.4× faster and token costs for generating document context fell by approximately 97%.
Visit Muro AI ↗
Joined 2025
Watch Your Health
Group Chief Technology Officer
After the acquisition of Rootally AI, I led a 50-person organization across engineering, data science, and product for a digital wellness platform serving more than 3 million active users.
I built the enterprise voice agent platform and oversaw product and platform architecture. Our team tripled release speed by improving deployment and engineering practices and reduced infrastructure costs by 25% through cloud migration, autoscaling, and optimization of machine learning workloads.
Acquisition announcement ↗
2020–2025
Rootally AI
Co-founder & CEO Singapore
I co-founded Rootally AI and led a 15-person engineering, clinical operations, research, and business team. We built phone-camera-based posture, fitness, and vital assessments for iOS, Android, and the web.
We secured contracts with insurers, employers, and hospitals. Rootally was among the top three solutions in Healthcare InnoMatch 2021, joined Iterative’s Winter 2024 cohort, and was acquired by Watch Your Health in 2025.
Feb 2021–Oct 2023
Shopee
Data Scientist, Sea Group Singapore
I developed and deployed intent recognition, prediction, ranking, and real-time query suggestions for Shopee’s support chatbot. We shipped to Singapore, Malaysia, Vietnam, Taiwan, the Philippines, Indonesia, and Brazil.
The intent models reduced support costs by over $5 million, with a 20% increase in chatbot resolution and a 30% increase in customer satisfaction. I also led development of conversational language models, built reusable machine learning services and APIs, and benchmarked LoRA fine-tuning across model sizes.
Jul 2020–Feb 2021
Cartrack
Data Scientist Singapore
I built real-time driver monitoring models using video and sensor data. I used multi-task learning to identify drowsiness and inattention and reduce false alerts.
Dec 2019–Jul 2020
ROSE Lab, NTU
Research Assistant Intern Singapore
At the Rapid-Rich Object Search Lab, I built real-time object detectors for NVIDIA Jetson TX2 and Xavier devices. I reduced model parameters and compute while maintaining accuracy and investigated class imbalance.
Mar–Aug 2019
I trained models to detect advertising policy violations, used active learning to select informative examples, and explored continual learning as data changed. I built an AWS pipeline for training and deployment and co-authored a BERT study for the AAAI Student Abstract Track.
Aug–Dec 2018
NVIDIA
Parallel Computing and AI Intern
On the RAPIDS team, I built a GPU system for training word2vec and worked on parallel block-circulant sparse matrix–vector multiplication to speed up computation and reduce neural network size.
May–Jul 2018
Mercedes-Benz R&D India
AI Intern
I prototyped computer vision algorithms that track the vehicle ahead to identify potholes and speed bumps with a single camera. We used road footage and simulation data and published the work as an SAE technical paper.