Looking for a job

Yatin Kande

@yatinkande

Message

Data scientist and ML engineer delivering measurable AI outcomes.

United States

Message

What I'm looking for

I seek roles building production ML systems where I can apply computer vision, RAG/LLM, and MLOps skills to deliver measurable business impact in collaborative, product-driven teams.

I am a Data Scientist and ML Engineer with hands-on experience building and deploying production ML systems across computer vision, generative AI, and predictive modeling. I have converted business problems into measurable AI outcomes across automotive, agricultural, and technology domains.

My work includes deploying YOLOv5 crop-disease detection at scale, improving minority-class recall with synthetic data, and delivering a churn model with 81% AUC-ROC that informed targeted retention outreach. I reduced inference latency by 40% through edge-compute and batching optimizations, and tracked model performance with dashboards to accelerate release cadence.

I design robust, production-ready architectures using cloud and infra tools such as AWS, Docker, Kubernetes, and MLflow, and I focus on reproducibility, monitoring, and measurable business impact while continuously iterating on model quality and deployment efficiency.

Experience

Work history, roles, and key accomplishments

Data Analyst Intern

DataZymes

Feb 2024 - Jul 2024 (5 months)

Deployed YOLOv5 crop disease detection on 10,000+ images achieving mAP 74% and F1 0.76, cut inference latency by 40% via batched processing and edge optimizations, and improved minority-class F1 from 0.72 to 0.76 with 1,500+ synthetic images to expand rare disease coverage.

Computer Vision Apache Spark OpenCV Edge Computing Power BI

Machine Learning Intern

SmartKnower

Mar 2022 - Jun 2022 (3 months)

Built an XGBoost churn model on 8,500+ records achieving 81% AUC-ROC, engineered 15+ behavioral features and applied SMOTE to improve minority-class detection, and optimized hyperparameters via GridSearchCV to increase precision while maintaining recall.