Part-Time Assistant Programmer

Penn State University • Ferguson Township, PA

25d • $57 • Remote

About The Position

The Department of Computer Science and Engineering within the College of Engineering is seeking applicants for a part-time programmer. The role: Help us build a focused, sharp system that turns peer-reviewed nutrition literature into transparent, evidence-based recommendations. You’ll implement the Extract, Transform Load (ETL) pipeline from OpenAlex, fine-tune and wire up transformer models for relation extraction, and ship a production-grade recommendation/generative layer backed by a provenance-rich knowledge graph.

Requirements

• Python 3.x
• modern NLP/ML (PyTorch or TensorFlow, Hugging Face, scikit-learn)
• data/ETL tooling (pandas, spaCy, Pydantic, Airflow/Prefect)
• APIs/services: RESTful design, auth, pagination, retries, structured logging, unit/integration tests, CI
• Storage/search: Postgres + vector/embedding store
• DevOps basics: containers, reproducible envs, simple local deploys, secrets management
• 3+ years professional Python and ML/NLP engineering, or equivalent portfolio
• Strong engineering hygiene (tests, docs, code review) and product sense
• Clear, direct communication in a remote team

Nice To Haves

• Deep NLP experience (relation extraction, weak supervision, prompt/adapter tuning).
• Semantic KGs (RDF/OWL, Neo4j/graph tooling), ontology work, and/or the Biolink Model.
• Experience with biomedical text corpora and literature mining.
• You’ve shipped scrappy, reliable research-to-prod systems before.
• graph DB experience welcome.
• Bonus: prior work at the intersection of ML + biosciences/clinical data.

Responsibilities

• Stand up and own a PubMed or OpenAlex ETL that ingests open-access articles, normalizes metadata, and keeps our corpus fresh.
• Use and/or fine tune transformer models (e.g., BERT variants) to extract semantic triples (like: (ingredient)–[biolink:ameliorates_condition]–>(health condition)); build evaluation and error-analysis loops.
• Implement a ranking layer that blends evidence strength (study design, sample size, effect size) with model confidence.
• Build the recommendation/generative service that balances constraints (evidence score, compatibility, regulatory limits) and exposes a clean API.
• Construct/maintain a knowledge graph that links every recommendation back to source papers (full provenance) and supports plain-language evidence summaries.
• Collaborate tightly with PI and domain scientists; deliver milestones on a fast, 12-month pilot timeline.

What This Job Offers

• Job Type: Part-time
• Career Level: Entry Level
• Education Level: No Education Listed
• Number of Employees: 5,001-10,000 employees