Open to opportunities

Farah Saeed

@farahsaeed

Doctoral-level machine learning researcher specializing in time-series foundation models and public-health applications.

Pakistan

Message

What I'm looking for

I seek research-driven roles applying foundation models and time-series forecasting to public-health or applied AI problems, in collaborative teams with open-source and publication opportunities.

I am a PhD-trained machine learning researcher focused on foundation models for time-series disease estimation, LLM fine-tuning, and applied deep learning for public-health outcomes. My work spans pretraining transformer-based networks, RAG for time-series, PEFT techniques (QDoRA, QLoRA), and applied projects like PeanutNeRF and robotic state-action recognition.

Previously I worked as a data scientist analyzing large-scale telecom datasets to produce customer segmentation, churn forecasting, and pipeline automation, and as an application engineer building domain-specific software. I contribute to open-source (Scikit-learn) and have multiple peer-reviewed and workshop publications applying ML to forecasting and computer vision.

Experience

Work history, roles, and key accomplishments

Doctoral Scholar

University of Georgia

Aug 2019 - Jul 2025 (5 years 11 months)

Researched foundation models and time-series disease estimation, pretraining PatchTST-based transformers and fine-tuning Llama/GPT variants with QDoRA/QLoRA to improve public-health forecasting and single-cell perturbation analysis.

Python PyTorch Hugging Face PEFT)PatchTST RAG Faiss Pgvector

Data Scientist

Jazz PMCL

Feb 2017 - Jul 2019 (2 years 5 months)

Performed advanced analytics on 40M+ telecom users to build behavioral profiles, churn prediction and engagement forecasting, and deployed Hadoop/Spark jobs to enable data-driven decisions.