Machine Learning Engineer for Geospatial AI Experimentation (Remote Sensing / Computer Vision / LLM / RAG) -- 2

We are conducting a research project in Geospatial Artificial Intelligence and Remote Sensing that supports an academic manuscript submission to journals such as:
ISPRS International Journal of Geo-Information
IEEE Journal of Selected Topics in Applied Earh Observations and Remote Sensing
The system integrates:
Satellite image processing
Computer vision object detection
Large Language Model (LLM) enrichment
Retrieval-Augmented Generation (RAG)
The engineering system already exists. Your role is to design and execute rigorous ML experiments and evaluation pipelines to support publication-quality results.
Infrastructure and deployment will be handled by a DevOps engineer.

System Architecture
The pipeline is based on the AWS open-source geospatial processing framework:
OSML ModelRunner
Reference deployment stack:
https://github.com/aws-solutions-library-samples/guidance-for-processing-overhead-imagery-on-aws
The system processes large satellite imagery through the following workflow:
Satellite imagery tiling
YOLO-based object detection
Detection clustering and filtering
LLM-based semantic enrichment
Optional RAG contextual reasoning

Three configurations will be evaluated:
CV-only baseline
CV + LLM enrichment
CV + LLM + RAG contextualization

Scope of Work
The hired ML engineer will design and execute the full experimental evaluation pipeline.
This project focuses on rigorous experimentation and evaluation, not developing new ML models.

Required Skills
Strong candidates should have experience in:
Machine Learning / Deep Learning
Computer Vision for remote sensing
Geospatial data processing
Experimental design for research papers

Technical stack:
Python
PyTorch or TensorFlow
OpenCV
NumPy / Pandas
Scikit-learn
Matplotlib / Seaborn
AWS cloud workflows
GPU computing

Experience with the following is highly desirable:
Satellite imagery processing
GeoJSON / GDAL
STAC catalogs
LLM integration
Retrieval-Augmented Generation

Nice-to-Have Experience
Experience with:
Academic ML research
Remote sensing journals
Geospatial AI
Multimodal AI pipelines
Experiment reproducibility frameworks

Deliverables Summary
The freelancer will produce:
Experiment protocol document
Dataset preparation pipeline
Controlled experiment runs
Evaluation metrics and statistics
Publication-ready figures and tables
Reproducibility documentation

DevOps engineers will handle:
AWS infrastructure
GPU environments
container deployment
storage configuration
The ML engineer focuses on experimentation and evaluation only.

NOTE: We have already selected the DevOps Engineer, Only ML is required.