Jay Vora

Data Scientist

Bengaluru, Karnataka, India5 yrs 8 mos experience

Highly StableAI Enabled

Key Highlights

3.5 years of experience in Computer Vision and Geospatial AI.
Proven ability to reduce turnaround time by over 90%.
Expertise in deploying ML solutions at scale using MLOps.

Stackforce AI infers this person is a Geospatial AI expert with strong capabilities in Computer Vision and Machine Learning.

Contact

Skills

Core Skills

Machine LearningComputer Vision

Other Skills

AWS SageMakerGeographic Information Systems (GIS)Python (Programming Language)LeadershipTeam ManagementDeep LearningData ScienceArtificial Intelligence (AI)OpenCVSAMGeospatial AIYOLOsegMask R-CNNImage ProcessingC++

About

Data Scientist with 3.5 years’ experience in Computer Vision, Geospatial AI, and ML model deployment. Skilled in PyTorch and AWS SageMaker with expertise in YOLO-based architectures, SAM, Mask R-CNN, and geospatial modeling. Proven ability to automate geospatial digitization (90%+ reduction in TAT) and deploy ML solutions at scale using MLOps practices (Docker, CI/CD pipelines, cloud infrastructure). Strong research foundation (IIT Kanpur M.Tech) with industry experience delivering production-ready ML solutions.

Experience

5 yrs 8 mos

Total Experience

2 yrs 7 mos

Average Tenure

5 mos

Current Experience

Ericsson

Data Scientist II

Dec 2025 – Present · 5 mos · Bengaluru, Karnataka, India · On-site

Aereo

Data Scientist

Jul 2022 – Nov 2025 · 3 yrs 4 mos · Bangalore · On-site

◦ Fine-tuned SAM (Segment Anything Model) with MANet on multi-modal remote sensing data to detect haul roads & benches in open-pit mines, reducing turnaround time by 95% and improving mine planning efficiency.
◦ Optimized YOLOseg on AWS SageMaker to detect rural features (roads, buildings, waterbodies), cutting digitization effort by 2–3 days per 10 sq. km village and enabling faster rural development planning.
◦ Applied Mask R-CNN on orthomosaic datasets for automated tree crown detection with 90% accuracy, supporting environmental monitoring projects, cutting survey costs by 40%+ and accelerating compliance reporting.
◦ Designed custom image processing algorithms for haul road and bench crest-toe detection using Digital Terrain Models, achieving 75% accuracy and reducing manual digitization effort by 70%.
◦ Led cross-functional research teams on geospatial ML use cases (building footprint detection, heap boundary extraction, point cloud classification), delivering scalable AI solutions to mining and infrastructure clients.

Computer VisionMachine LearningAWS SageMakerGeographic Information Systems (GIS)Python (Programming Language)Leadership+5