Teshome A.

Machine Learning Engineer specializing in crop yield prediction, EO data, and scalable ML pipelines. I’m a Machine Learning Engineer specializing in building end-to-end data pipelines and predictive models for real-world applications, with strong experience in crop yield prediction, document intelligence, and multi-source data integration. I help clients turn complex, messy data into accurate predictions, automated systems, and scalable solutions. 🔹 What I do best: Build machine learning models (XGBoost, Random Forest, Deep Learning) Design data pipelines (ETL, preprocessing, feature engineering) Work with Earth Observation (EO) data (Sentinel-2, NDVI, EVI) Develop document AI systems (OCR, AWS Textract, structured data extraction) Deploy solutions using FastAPI, APIs, and production-ready workflows 🔹 Recent impactful work: Led crop yield prediction modeling using EO + environmental + ground truth data (ViSTA Data Sprint) Built AWB document processing systems extracting structured data with high accuracy Developed automated pipelines integrating OCR, validation, and DMS upload Designed ML deployment systems with API + dashboard integration 🔹 Advanced / Foundational Model Work: Worked with foundational EO models (e.g., Panoptic/Prithvi-style models) to generate rich imagery-based features aligned with ground truth agricultural data Extracted high-dimensional representations from satellite imagery to enhance crop yield prediction performance Integrated foundation model embeddings + environmental + EO features into ML pipelines for improved generalization Explored transfer learning approaches, reducing the need for training models from scratch 🔹 What you can expect: Clean, scalable, and production-ready solutions Strong problem-solving and attention to detail Clear communication and reliable delivery