Kalpesh M.

Key Skills

Software

Other

Top Subject Matter

Bioimaging Domain Expertise

Microscopy Domain Expertise

multispectral health diagnostics

Top Data Types

Image

3D Sensor

Document

Top Task Types

Segmentation

Classification

Tracking

Text Generation

Freelancer Overview

Research Manager, Machine Learning Group - Dyson. Brings 15+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other, Internal, and Proprietary Tooling. Education includes Chief Technology Officer Programme, National University of Singapore (2025) and Doctor of Philosophy, National University of Singapore (2012). AI-training focus includes data types such as Medical, DICOM, and Image and labeling workflows including Segmentation, Classification, and Tracking.

Labeling Experience

Founder & CEO | Navi-Vis Pte. Ltd., Singapore (June 2025 – Present)

DocumentText Generation

Built an AI-driven platform that transforms fragmented technical, clinical, and design evidence into pre-submission-ready regulatory documentation. The workflow ingests PDFs and related artifacts to support structured gap analysis and draft preparation, which requires labeling/structuring evidence and generating compliant text outputs. Defined a human-in-the-loop regulatory process to refine evidence traceability and draft generation aligned with regulatory guidance.• Ingests PDFs/images/design/validation artifacts to produce structured regulatory documentation.• Performs structured gap analysis that depends on labeled evidence spans and categories.• Uses human-in-the-loop review to improve the quality of generated draft text.• Develops evidence traceability and drafting aligned with regulatory guidance for submissions.

2025 - Present

Founder-in-Residence | Singapore Deep-Tech Alliance (SDTA) (Jan 2025 – Present)

Other3D SensorClassification

Led end-to-end solution development for projects combining 3D image sensor data and ultrasound imaging to improve automation and data collection efficiency. Work on sensor- and imaging-driven solutions typically requires creation or preparation of labeled training/evaluation data to validate automated detection and analytics. Translated imaging and AI concepts into commercialization pathways with an emphasis on clinically relevant solutions.• Developed solution concepts for imaging data pipelines intended for model training and evaluation.• Supported integration of 3D sensor and ultrasound data workflows for automated data collection.• Worked on clinical imaging/AI commercialization paths requiring validated outputs.• Coordinated project execution across venture-building initiatives centered on imaging AI models.

2025 - Present

Research Manager, Machine Learning Group - Dyson

ImageSegmentation

Spearheaded product strategy and technical roadmap development for AI-enabled personal care diagnostics integrating sensing, imaging, and biomarker analytics. Set up multi-million-dollar lab infrastructure with research institutes to accelerate R&D and clinical evidence generation. Directed cross-functional programs across microscopy, imaging, time-series, and unstructured data while leading specialists in bioimaging, data engineering, and machine learning. • Led full-cycle innovation from roadmap to evidence generation • Built biomarker extraction and privacy-preserving anonymization workflows • Developed microscopy segmentation and drift correction using advanced models • Supported clinical study planning including sample size estimation and statistics

2021 - 2024

Research Manager, Machine Learning Group, Chief Research Office Group | Dyson (Apr 2021 – Nov 2024)

OtherSegmentation

Built and improved high-precision microscopy segmentation workflows using Segment Anything Models (SAM) and developed drift detection and correction algorithms for consistent model performance over time. Supported clinical sample size estimation and statistical planning for multisite studies to make imaging analysis outputs suitable for downstream evaluation. Worked on end-to-end pipelines where ground-truth/label quality was critical for training and validation of biomarker extraction models.• Used segmentation outputs from microscopy data to support health biomarker extraction from multispectral imagery.• Implemented drift detection/correction to maintain labeling consistency across imaging sessions.• Participated in statistical planning and study design to support supervised learning and validation.• Coordinated with specialists in bioimaging, data engineering, and machine learning for dataset readiness.

2021 - 2024

Senior Physicist, Advanced Technology Group - Applied Materials

ImageSegmentationTracking

Defined minimum viable products and commercial system requirements in coordination with internal stakeholders and academic collaborators. Co-supervised a postdoctoral researcher in a joint lab developing machine-learning-based image focus tracking for high-throughput microscopy platforms. Developed hardware drivers, automation algorithms, image stitching and alignment pipelines, and quality assurance methods for high-resolution microscopy systems. • Translated technical needs into system requirements and MVP scope • Advanced ML-based focus tracking for imaging platforms • Engineered automation and image processing pipelines • Contributed to segmentation and bioinformatics workflows for spatial transcriptomics

2018 - 2021

Show more

Education

Work History