Materials Domain Specialist (Knowledge Architect)

Materials Domain Specialist (Knowledge Architect)

Patsnap's Materials team is building a connected, continuously improving corpus of structured materials knowledge, drawn from patents, scientific literature, technical datasheets, standards, and other sources of materials data. That corpus powers AI products that help R&D and IP professionals search, extract, and reason over materials science, and it is increasingly the ground truth that our own AI agents reason with.

You will be the architect of that understanding. You decide how materials knowledge is represented: how a composition, a property, a processing condition, and a measurement are modelled so that a claim in a patent, a table in a paper, and a value on a supplier datasheet become comparable, connected, and verifiable facts about the same material. You define what counts as evidence, how confidence and provenance travel with every fact, and how the corpus learns from its own use.

You will set the scientific and evidential standard for the corpus: what a source explicitly discloses, what it merely permits or implies, and whether a property or measurement is attached to the correct material, formulation, or tested sample. Collaborating with data scientists and engineers, you will turn that judgment into schemas, ontologies, reproducible guidelines, and benchmarks. You will also participate in the design of the feedback loops that let humans and agents make the corpus better every time they use it.

This position is based in our Singapore office.

Want to see the platform you'd be representing?
Check out this short overview:

This is an in-office position based in our Singapore office.

 

Who are we?

Patsnap is a global, pre-IPO company that transforms the way organizations harness their Intellectual Property and Research & Development productivity. Our platform revolutionizes how IP and R&D teams collaborate across the entire innovation lifecycle, using domain-specific AI to accelerate the creation of market-ready products. With over 18,000 customers worldwide, including some of the biggest names in innovation, Patsnap is at the forefront of technological advancement.

We have a vibrant and diverse team with offices in Singapore, Toronto, London, Shanghai and remote teams based in US. Our hyper-growth trajectory is powered by our people, and we are extremely proud of our company-wide vision, work ethic, and entrepreneurial spirit. We are committed to fostering an inclusive environment where talent thrives and ideas bloom.

What You'll Be Doing:

• Architect the materials knowledge model: own the ontology, schema, and normalization layer that represents materials, compositions, structures, properties, processing, and measurement context, and design it to hold across patents, literature, datasheets, and other sources while preserving each source's disclosure semantics.

• Define the evidence standard: set the rules for what a source discloses versus implies, how facts are linked to the correct material, formulation, or tested sample, and how provenance, confidence, and conflicting values are represented so that both humans and AI agents can trust and trace every fact.

• Connect the corpus: design how entities and facts from different sources resolve to one another (the same polymer grade in a patent, a paper, and a supplier catalogue), and how relationships across materials, applications, and performance are made explicit and queryable.

• Build the benchmarks and gold standards that measure extraction, search, and agentic reasoning against the corpus: define task scope, adjudicate gold sets, set metrics and release gates, and maintain regression cases.

• Make the corpus self-improving: work with engineers and product to close the loop from customer usage, agent traces, and observed model failures back into versioned schema changes, rules, and curation fixes, with measurable gains in coverage, quality, speed, and cost.

• Run internal and external curation operations, including reviewer calibration, vendor delivery, quality gates, and throughput and cost reporting.

• Serve as the team's lead adjudicator for difficult materials-data questions, and be the domain voice in decisions about how agents should query, weigh, and reason over materials knowledge.

Why This Role：

• Work at the frontier of what LLMs can reason about, and push past it. Frontier models still confuse a claimed range with a measured value, a comonomer with a blend, a datasheet grade with the polymer family it belongs to. The structured knowledge you build is what lets an agent get these right when the model alone cannot.

• See your judgment scale. Every distinction you formalise, from disclosure versus inference to property-to-sample attribution, becomes a rule that runs across millions of documents and every agent decision, and every agent failure comes back to you as the next thing to formalise.

• Own the capability that decides what Patsnap's materials AI can know, verify, and defend, at the centre of the materials roadmap and with a direct line into how our agents are built.

Core Requirements:

• Degree in polymer science, chemistry, materials science, chemical engineering, metallurgy, or a related field; or equivalent professional experience.

• 5+ years working deeply with materials or chemical technical information, for example examining and analyzing patents, curating scientific databases, building materials data resources and/or product master data management.

• Strong evidence judgment across source types: able to distinguish claimed scope from example-level facts, explicit disclosure from inference, and reported values from their measurement conditions, in patents as well as papers and datasheets.

• Able to interpret polymer structures, repeat units, generic or Markush definitions, formulations, ranges, measurement context, and property-to-sample relationships.

• Experience designing or stewarding structured representations of domain knowledge: ontologies, schemas, taxonomies, controlled vocabularies, or normalization rules.

• Experience translating expert judgment into repeatable guidelines and deterministic rules.

• Experience leading projects, reviewers, or specialist partners against delivery schedules and acceptance criteria.

• Able to communicate complex decisions and concepts clearly to product, and engineering teams.

• Professional working proficiency in English and Chinese.

Nice-to-haves (not required — you'll have room and support to pick these up on the job):

• Knowledge graph, entity resolution, information retrieval, or search-evaluation experience.

• Familiarity with how LLM-based agents retrieve and use structured knowledge (retrieval-augmented generation, tool use, evaluation of agent outputs).

• Chemical structure, cheminformatics, or scientific-database experience.

• Experience managing outsourced or distributed teams.

• Proficiency with agentic productivity tools (e.g. Claude Code, Codex).

• Professional proficiency in an additional patent language.