Sarthak Jain Resume
Sarthak Jain Resume
PROFESSIONAL SUMMARY
Data Scientist at SatSure with expertise in machine learning, deep learning, and data engineering. Skilled in
developing AI-driven solutions, optimizing data pipelines, and deploying scalable models. Passionate about
leveraging AI to drive automation and enhance decision-making processes.
TECHNICAL SKILLS
WORK EXPERIENCE
● Developed and implemented machine learning and deep learning models, utilizing computer vision
and segmentation techniques, to address diverse problem domains.
● Applied machine learning-driven time series analysis, utilizing models such as Random Forest (RF),
Long Short-Term Memory (LSTM), and others, to conduct pixel-level analysis of geospatial temporal
data, to generate actionable insights for data-driven decision-making.
● Developed a deep learning-based classification model using U-Net in PyTorch with Sentinel-1 &
Sentinel-2 data, achieving an F1 score of ~0.85 across multiple crops and reducing manual classification
efforts by 250+ hours through a Dockerized deployment.
● Developed a captcha-cracking automation using OpenCV, CNN-based models, and OCR techniques,
reducing data fetching time by ~25%.
● Designed a segmentation-based farm boundary detection model in PyTorch using Sentinel-2 data,
achieving a dice score of ~0.88 (India) and ~0.94 (global), and deployed it as a scalable, Dockerized
product.
● Engineered an automated validation pipeline with statistical integrity checks, improving data accuracy
by 40% and integrating results into an intuitive Tableau dashboard.
● Optimized data pipelines for memory efficiency (70-80% reduction) and processing speed, cutting
execution time by 50% for large-scale geospatial datasets.
● Built interactive Tableau dashboards to visualize key metrics, enhancing data-driven decision-making
for cross-functional teams.
● Conducted in-depth literature reviews and research-driven tasks to support data science initiatives
and model improvements.
PUBLICATIONS
Mitigating Bad Ground Truth in Supervised Machine Learning based Crop Classification: A Multi-Level
Framework with Sentinel-2 Images [ Publication Link]
EDUCATION
INTERESTS
● Performed stand-up comedy, blending humor with corporate life experiences for office events.
● Regularly play badminton, promoting physical fitness and teamwork.