Ref Id: HREZ-12345
Published Date: Jun 1, 2026
Job location: Brazil, Colombia, Costa Rica
General information
Description
HATCHWORKS AI Senior Knowledge Engineer IC3–IC4 · SHACL / RDF · Semantic Web · Remote / Costa Rica, Brazil, Colombia |
Remote / LatAm | 5+ Yrs Exp. | English Required | Full-Time |
THE ROLE
Most engineers build software. Few build the rules that govern it — and almost none make those rules machine-checkable.
We're looking for a Senior Knowledge Engineer who lives at the intersection of semantic web, domain modeling, and production engineering. You will extend production knowledge graphs at major enterprise clients, encode complex domain rules as SHACL shapes, and deliver explanation layers that turn structured violations into plain-language rationale users can act on.
This is one of the highest-leverage roles on the engagement. The rules you author gate every recommendation the system makes. The graph you build is the substrate every other engine queries.
If you've spent years explaining what SHACL is to colleagues — come work somewhere it's the spine.
WHAT YOU'LL ACTUALLY DO
- Extend production RDF graphs — walk existing graphs line-by-line with client architects, understand their model, propose targeted extensions rather than rebuilds.
- Author SHACL eligibility rules — domain compliance, product-quality, regulatory constraints — co-authored with SMEs from specification documents.
- Build eligibility APIs that return structured verdicts (eligible / not, plus reasons) with sub-second latency, called per-decision by upstream optimization and forecasting systems.
- Co-design the eligibility-query interface with the engineers consuming it — the contract between rules engine and solver is a joint design artifact.
- Implement explanation layers on top of structured violations — turn raw rule failures into plain-language rationale users can read in seconds.
- Apply generative AI patterns thoughtfully — RAG-grounded explanation generation, structured-to-natural-language layering, small-model orchestration where it fits the explanation problem.
- Partner with domain SMEs to extract tribal knowledge and codify it. The output is a tested, versioned rule set — not a wiki.
- Pair with client engineers embedded in the team — knowledge engineering is a rare discipline; building expertise inside the client team is part of the role.
- Design SME-friendly rule authoring workflows — so the rule set becomes maintainable by domain experts after the engagement ends.
- Validate rules against production data — test cases drawn from real entities, real transactions, real edge cases.
- Contribute to HatchWorks AI's knowledge engineering practice as a discipline that's still being defined industry-wide.
THE TECHNICAL BAR
You will be expected to execute hands-on technical work from day one. The requirements below reflect the actual skills needed to deliver outcomes for enterprise clients.
Must-Haves | |
Semantic Web | 5+ years — RDF, OWL, SPARQL in production contexts |
SHACL | 3+ years authoring SHACL shapes; understanding of advanced features (rules, paths, targets) |
Python | Production-quality; familiarity with PySHACL or equivalent |
Triplestores | Stardog, GraphDB, Blazegraph, or equivalent — production experience |
Domain Modeling | Translating SME knowledge into formal logic; ontology design fundamentals |
API Design | Sub-second eligibility services; production performance tuning |
Generative AI Patterns | RAG, structured-to-natural-language explanation, small-model orchestration |
SME Collaboration | Documented experience working with non-technical domain experts |
Cloud | Production experience with a major cloud platform |
CORE TECH STACK
Full Stack | |
Languages | Python · SPARQL · Turtle / RDF/XML |
Graph Databases | Stardog · GraphDB · Apache Jena · Blazegraph · Neo4j (a plus) |
Validation | PySHACL · TopBraid · SHACL Playground |
Explanation | RAG frameworks · LLM-augmented explanation layers · Custom NL generation |
Cloud | Major cloud platforms · Lakehouse environments |
DevOps | Docker · CI/CD · Git |
NICE TO HAVE
- Regulated-industry experience — domains with formal compliance rules
- PROV-O, SKOS, or other upper-ontology experience
- Domain-specific languages for business rules — Drools, Datalog
- Production experience with reasoners — Pellet, HermiT, ELK
- Publication or open-source contribution in semantic web
WORKING CONTEXT
This is an active, client-embedded delivery engagement — not a research role. You will ship production rule systems from week one.
All team members operate within client timezone hours. LatAm candidates are strongly preferred.
Semantic web is a niche skill. HatchWorks AI takes it seriously — this is a first-class discipline, not a side project. You will have peers who speak the language.
THE HUMAN BAR
Technical depth is rare. Translator-of-domain-knowledge is rarer.
The engineers who thrive here can sit in a room with a domain SME, walk through a specification document, and emerge with a SHACL shape that captures the rule, plus a test case the SME signs off on. That's the job — not writing Turtle for its own sake.
- SME-first orientation — you treat domain experts as the source of truth. You extract their knowledge by listening, not by lecturing about ontologies.
- Explanation as a deliverable — a rule that fires correctly but can't explain itself is incomplete. Plain-language rationale is part of the work.
- Pragmatic semantic web — you know when to use formal inference and when to skip it. You don't reach for OWL DL when SHACL closes the case.
- Maintainability mindset — you design rule sets that survive your handoff. SME-authorable workflows are a feature, not an afterthought.
WHAT MAKES THIS DIFFERENT
Knowledge engineering roles are rare. Production knowledge engineering roles are rarer still. Most companies treat semantic web as research or as a graph-database integration. We treat it as the substrate every decision runs on.
You will be embedded with a major global enterprise client — operating in supply chain, logistics, manufacturing, or adjacent operational industries — building production AI systems that decision-makers actually use.
Your rules gate every downstream recommendation. Your explanations are what a user reads when they ask why an action was or wasn't recommended. The work is high-leverage, highly visible, and deeply embedded in production decision flow.
YOU'LL WORK INSIDE A PROPRIETARY AI DELIVERY FRAMEWORK
Most teams using AI tools today are flying blind — they feel faster but can't prove it. We built GenDD (Generative Driven Development) to change that, and you'll be working inside it from day one.
GenDD is a proprietary methodology and toolset built around one belief: great engineers need full context to move fast without breaking things. So we built the tools to give them exactly that.
Zero Ramp-Up | Proven Value | All-In Team |
Our tooling surfaces full project context — architecture, data flows, history — so you hit the ground running on every engagement. | GenDD instruments every agent interaction, commit, and deploy — turning your work into clear, client-ready evidence of real impact. | Everyone here uses agents seriously. You're not evangelizing — you're exploring. We learn fast, share everything, and enjoy the work. |
This is the environment technologists thrive in: structured methodology, tooling that solves real problems, teammates who move at pace, and clients who care about the results.
HOW TO APPLY
Send us three things:
1. Why this role specifically — A brief note (3–5 sentences) on what draws you to this engagement and technical area. Be specific about the programme context and your relevant experience.
2. Something relevant you've delivered — A concrete example of work you've produced in this domain — a project outcome, a script, a report, a case study. We want evidence, not claims.
3. Your resumé or LinkedIn profile.
We read every application. We respond to the ones that show us you've actually done this.
We are an equal opportunity employer. We evaluate candidates on technical merit, communication, and demonstrated capability — not credentials, geography, or background.
< Back to job list
Senior Knowledge Engineer application
