À propos de ce poste Materials & Polymers Domain Specialist chez Patsnap
About the role
Patsnap is building evidence-backed materials datasets extracted from patents and scientific literature — starting with an engineering-polymer curation program. You will own the delivery of this program end to end: managing any vendor relationships, safeguarding scientific quality, and connecting the curated output to Patsnap's ontology, extraction, and search teams. Beyond this program, you will support a growing portfolio of materials data extraction, indexing, and search projects.
This is a hands-on scientific role, not a pure project-management role. The hardest problems are chemical, not infrastructural: deciding whether a Markush structure's variable definitions were captured faithfully, whether a repeat unit was explicitly disclosed or inferred by a curator, whether a copolymer's architecture was stated or assumed, whether a property was bound to the right sample among a table of examples. You will be the final scientific arbiter on these calls, and you must be able to make them yourself.
What you will do
- Run the curation program. Own annotation guidelines, calibration rounds, delivery schedules, and acceptance criteria with internal & external data curators. Track throughput, cost, and error rates by task type and report against pilot go/no-go criteria.
- Own scientific quality. Design and run sampling-based QC on curated polymer data: gold-standard adjudication, inter-annotator agreement, held-out precision and coverage measurement, and independent review of hard cases (generic/Markush structures, repeat units, composition and formulation tables, measurement context). Evaluate AI/ML pre-annotation and extraction outputs against curated gold standards, and turn error analysis into improved guidelines and dictionary rules.
- Steward the data contracts. Manage evolution of the evidence, composition, and normalization-dictionary contracts; review normalization rules and dictionary releases; triage unresolved-term and ontology-gap queues with Patsnap's ontology team.
- Connect data to product. Work with search and product teams to translate curated relationships into indexing and multi-hop query capability; maintain query benchmarks that demonstrate downstream lift.
- Support the broader portfolio. Apply the same delivery and QC discipline to other materials data extraction, indexing, and search projects as the program expands beyond polymers.
Core requirements
These three are non-negotiable; everything else is trainable.
Polymer science depth. Masters in polymer science, chemistry, materials science, or chemical engineering — or a Bachelors with equivalent industry experience — with command of polymer structure and representation, demonstrated by the ability to:
-
Quality by process. Demonstrated experience managing quality through structured process — lab quality systems (GLP/ISO), audit readiness, structured review workflows, or data QC by sampling and metrics — rather than ad-hoc checking.
-
Delivery management. Experience managing projects, teams, or external partners against defined deliverables, schedules, and acceptance criteria.
You likely come from one of these backgrounds
We are deliberately recruiting from several adjacent pools. Strength in one of these, plus the core requirements above, is a complete application — you do not need to tick every box.
- Industry polymer/materials scientist moving toward data: formulation, adhesives, coatings, or specialty-chemicals R&D, ideally with exposure to digitalization, master data, or lab-informatics initiatives.
- Patent professional with chemistry depth: patent analyst, searcher, examiner, or IP-firm scientist specializing in polymers or chemistry.
- Scientific content/curation operations: curation, content QC, taxonomy, or ontology work at a scientific database or content provider — provided you also have genuine chemistry or materials depth, not solely content-workflow experience.
- Cheminformatics or materials informatics: research institute or industry roles in materials data, structure representation, or chemical databases.
We will train you on
Nice-to-have
- Applied AI/ML literacy: evaluating LLM or ML extraction output (error analysis, precision/recall, gold-standard evaluation design) or working in AI-assisted annotation workflows. Model building is not required — that stays with our engineering teams.
- Experience managing outsourced or distributed annotation/curation teams.
- Knowledge-graph or ontology experience (property graphs, SKOS/OWL, entity resolution).
- Scripting ability (Python or similar) for validation, sampling, and QC automation.
- Familiarity with information retrieval or search-relevance evaluation.
- Cheminformatics tooling (structure representations, SMILES/InChI, structure normalization).
