Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I’m Alexander Amiri, a data-focused AI practitioner who enjoys turning messy, real-world datasets into reliable pipelines and decision-ready insights. I’ve worked on everything from integrating multi-carrier shipment data into centralized systems to building LLM workflows that summarize clinical notes efficiently for downstream biomedical tasks. In my recent roles, I’ve developed automation that reduces manual operational work, created exception-focused reporting so non-technical teams can self-serve account history, and produced large-scale MRI–pathology training datasets. I’m especially interested in practical ML/LLM engineering—ETL, MLOps, evaluation, and fine-tuning—so models can run accurately at scale and support healthcare and operational decision-making.…I’m Alexander Amiri, a data-focused AI practitioner who enjoys turning messy, real-world datasets into reliable pipelines and decision-ready insights. I’ve worked on everything from integrating multi-carrier shipment data into centralized systems to building LLM workflows that summarize clinical notes efficiently for downstream biomedical tasks. In my recent roles, I’ve developed automation that reduces manual operational work, created exception-focused reporting so non-technical teams can self-serve account history, and produced large-scale MRI–pathology training datasets. I’m especially interested in practical ML/LLM engineering—ETL, MLOps, evaluation, and fine-tuning—so models can run accurately at scale and support healthcare and operational decision-making.

Alexander Amiri

Data Scientist, AI Engineer, Data Analyst, +2





I’m Alexander Amiri, a data-focused AI practitioner who enjoys turning messy, real-world datasets into reliable pipelines and decision-ready insights. I’ve worked on everything from integrating multi-carrier shipment data into centralized systems to building LLM workflows that summarize clinical notes efficiently for downstream biomedical tasks. In my recent roles, I’ve developed automation that reduces manual operational work, created exception-focused reporting so non-technical teams can self-serve account history, and produced large-scale MRI–pathology training datasets. I’m especially interested in practical ML/LLM engineering—ETL, MLOps, evaluation, and fine-tuning—so models can run accurately at scale and support healthcare and operational decision-making.…I’m Alexander Amiri, a data-focused AI practitioner who enjoys turning messy, real-world datasets into reliable pipelines and decision-ready insights. I’ve worked on everything from integrating multi-carrier shipment data into centralized systems to building LLM workflows that summarize clinical notes efficiently for downstream biomedical tasks. In my recent roles, I’ve developed automation that reduces manual operational work, created exception-focused reporting so non-technical teams can self-serve account history, and produced large-scale MRI–pathology training datasets. I’m especially interested in practical ML/LLM engineering—ETL, MLOps, evaluation, and fine-tuning—so models can run accurately at scale and support healthcare and operational decision-making.

Available to hire

I’m Alexander Amiri, a data-focused AI practitioner who enjoys turning messy, real-world datasets into reliable pipelines and decision-ready insights. I’ve worked on everything from integrating multi-carrier shipment data into centralized systems to building LLM workflows that summarize clinical notes efficiently for downstream biomedical tasks.

In my recent roles, I’ve developed automation that reduces manual operational work, created exception-focused reporting so non-technical teams can self-serve account history, and produced large-scale MRI–pathology training datasets. I’m especially interested in practical ML/LLM engineering—ETL, MLOps, evaluation, and fine-tuning—so models can run accurately at scale and support healthcare and operational decision-making.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Work Experience

Data Scientist at Hedwig Xpress LLC

January 1, 2026 - Present

Consolidated shipment records and delivery history across 3+ carrier integrations into a centralized MongoDB database, eliminating manual spreadsheet tracking and reducing data retrieval time by 65%. Automated weekly customer order ingestion using a Python pipeline, cutting manual data entry overhead by 70% and standardizing records across 500+ monthly shipments while generating $4.5K/month in operational savings. Developed delivery reporting views that enable the operations team to independently query account history and flag exceptions, reducing reliance on technical staff for routine data lookups.

AI Engineer at UCSF Radiation Oncology

October 1, 2025 - Present

Built an LLM summarization pipeline (Qwen2.5/Ollama) that compresses 100K+ token clinical notes to ~500 tokens, enabling BioClinicalBERT ingestion across 7 glioma follow-up timepoints. Processed large-scale clinical datasets including 64K MRI studies and 26K pathology reports across 12K patients, producing 50K high-quality MRI–pathology training pairs for downstream classification tasks. Fine-tuned encoder LLMs using LoRA and contrastive learning (MNRL), improving matched-pair cosine similarity from 0.63 to 0.79 across 70K patient-timepoint pairs with zero-error processing.

Data Analyst at UC Davis Lebril la Lab (Clinical Glycomics & Cancer Biomarker Research)

March 1, 2023 - April 1, 2024

Developed an ML pipeline for automated LC-MS annotation, implementing a kernel Naive Bayes classifier achieving 82% accuracy in glycan structure prediction from mass spectrometry data. Analyzed clinical samples from 50,000+ patient cohorts to identify cancer-associated glycan biomarkers, processing mass spectrometry data using Python and SQL. Synthesized 20+ peer-reviewed publications on glycomics and biomarker discovery and presented findings at bi-weekly lab meetings within a 10-person multidisciplinary research team.