Anwar A.

Project Gaia:

I worked on integrating and analyzing ecological data using machine learning and deep learning models. The project focused on leveraging Python and Azure OpenAI to explore generative AI (GenAI) use cases, particularly through prompt engineering. By developing custom prompts and fine-tuning pre-trained GenAI models, we aimed to derive meaningful insights and predictions regarding biodiversity, climate change, and species distribution. Additionally, I contributed to building and optimizing the data pipeline, ensuring efficient processing of large ecological datasets. The combination of machine learning algorithms and prompt engineering enabled us to create innovative tools for ecological research, enhancing decision-making for biodiversity conservation and environmental studies.

Project Grace:

I played a pivotal role in developing a NoSQL database solution for carbon simulation models, designed to assist consulting firms in their environmental sustainability efforts. I built a robust ETL pipeline that streamlined data ingestion, transformation, and storage. The project utilized Azure CosmosDB and MongoDB to manage large volumes of unstructured and semi-structured data, ensuring flexibility and scalability for future data growth. The ETL processes facilitated seamless integration of diverse data sources, enabling real-time analysis and visualization of carbon footprints. This work was crucial in providing actionable insights to help businesses reduce their environmental impact and optimize their sustainability strategies.

Internal Projects: I contributed to the development of two major internal projects:

Photo Sharing Software: Designed and implemented a software solution for efficient photo sharing within a cloud-based architecture. This project involved configuring and deploying services using AWS CloudFormation for infrastructure as code, Amazon S3 for scalable storage, and DynamoDB for a high-performance NoSQL database. The project utilized Python for backend development and Git for version control, ensuring seamless team collaboration.

Predictive Finance Project: I played a key role in building a predictive analytics platform for financial forecasting. The project involved creating a serverless architecture using AWS SAM and integrating various AWS services like S3 and DynamoDB for data storage and retrieval. Docker was used for containerized deployment, and OpenBB was leveraged for financial data analysis and visualization. Python was the primary programming language used to implement data processing pipelines and predictive algorithms.

Client Project: I led the development of a sophisticated data pipeline to automate table recognition and extraction from large volumes of documents for a client. This project utilized the machine learning capabilities of Amazon SageMaker to train and deploy models for accurate table extraction. Python was used to build the data processing logic, while AWS Lambda was employed to create serverless functions for seamless integration and scalability.

After my engineering degree, I chose to deepen my knowledge by pursuing a PhD. Here is the abstract: The study of biology combines shape studies (diversity) and process modeling (functional or evolutionary). Pleiade addresses the dual challenge of rapid and relevant measurement of dissimilarities between biological objects and the exploration of relationships between trait diversity and function diversity, at multiple scales. We are developing algorithms, models, and software frameworks for applications in ecology, evolution, and biotechnology. This is a thesis offer in the context of a collaboration between Inra and Inria, combining expertise in ecology, statistical modeling, and high-performance computing. The goal is to build OTUs (Operational Taxonomic Units) using metabarcoding tools (and evolving them) from protist community datasets from NGS (Next Generation Sequencing), with the thesis challenge of scaling up statistical methods given the size of the datasets. The expected work is therefore multidisciplinary, with a basic disciplinary competence in statistical modeling and high-performance computing. The candidate should show an openness to the applied aspects of these disciplines, particularly in digital ecology and taxonomy. The ability to work in a multidisciplinary context is both an asset of this position and a required skill: statistical modeling, metabarcoding and data analysis, and HPC.