Data Engineering Intern

**About us:** **Novobiom** is a fast-growing impact start-up harnessing fungi's remarkable degradation powers to treat waste and decontaminate polluted sites. At the computational biology unit we are building a digital platform for the **spatiotemporal simulation of microbial communities** — fungi and bacteria working together to decontaminate polluted land sites (crude oil products, PFAS, DDT, …) and restore soil health. The long-term goal is a fast, predictive simulator that bridges genome-scale metabolic models (GEMs) with AI surrogates, validated against in vivo and in vitro data. Over the years, our experimental work has generated a large volume of valuable data — substrate concentrations, biomass/OD curves, CO₂ traces, community compositions, sequencing outputs — currently scattered across reports, subreports and loose Excel tables. Our modeling pipeline has reached the point where the binding constraint is no longer modeling talent or compute: it is structured, linked, experimentally-grounded data. **Role overview & Responsibilities:** Your mission is to gather all of the company's experimental data and turn it into one coherent, modeling-ready asset: **structured** (a consistent schema, not a pile of tables), **linked** (records connected along three axes — environment composition, community composition, and time — so a query can trace how an environment and its community co-evolved), and **canonical** (every metabolite, organism and condition mapped onto our existing modeling namespace, so the database plugs directly into our simulation stack). A key enabler is modern AI tooling (e.g. Claude Code), which makes it feasible for one person to sweep an entire documentation corpus and extract structured records within an internship window. The AI does the volume; you own the correctness — designing the extraction, enforcing the schema, and verifying values with provenance and confidence metadata attached to every record. The work unfolds in three overlapping phases: * **Source inventory & extraction** — catalogue every data-bearing artifact (reports, Excel tables, bioreactor logs, sequencing outputs, lab notebooks) and extract structured experiment → environment → community → measurement → timepoint records with provenance attached. * **Schema design & canonical linking** — design the database around the environment – community – time axes; map everything onto a canonical namespace (e.g. MetaNetX harmonisation); resolve units, duplicates and ambiguities. * **Analysis & coverage map** — extract first insights from the assembled data (uptake/growth dynamics per organism and substrate, community interactions, time-course behaviour, metabolite concentrations, …) and produce a coverage-and-gaps map showing where future experiments would most cheaply pay off. Deliverables include a unified, queryable database; a reusable ingestion pipeline that keeps absorbing new reports as they are produced; and an analysis notebook with first observations on dynamics and interactions. If the core phases land well, natural stretch goals include a first exploratory fit of an environment → uptake predictor on the newly-structured data, and hardening the ingestion pipeline for continuous automated use. **Qualifications:** Curious students (Master students preferred) from any of: bioinformatics, computational biology, biology, biotech, physics, computer science, data science, applied math. Deep ML or metabolic-modeling expertise is not required — the biology context can be learned on the job. What matters most is cross-disciplinary curiosity and learning agility: the ability to move between messy biological documents, data engineering, and modeling priorities without getting stuck at any one of them. **Required:** * Comfortable with Python data wrangling (pandas) * A sense for data/database design * Fluency with (or eagerness to learn) AI tooling — and the judgment to distrust it in the right places **Plus:** * Database technologies — SQL, schema design, data validation * Scientific computing — Git, OOP, cloud workflows * Microbiology / biochemistry literacy **Double plus:** * Prior exposure to metabolic modeling, flux balance analysis (FBA) and/or genome-scale metabolic models (COBRApy or similar) * Experience wrangling experimental biological data (growth curves, omics outputs, bioreactor logs) Send a short email — a paragraph about you, your experience related to the responsibilities and qualifications, and your CV. **[email protected]**