Free-form CVs and vocational certificates contain rich latent signals that conventional keyword matching routinely fails to capture. We develop an entity-extraction and graph-alignment architecture that maps natural language experience to standardized international taxonomies (ESCO/OFO) with calibrated confidence, separating job-title marketing from genuine operational mastery.
Can structural competency graphs recover genuine practitioner capabilities without succumbing to job-title ambiguity?
In our parsing experiments over 4,800 vocational CVs, words like "managed" or "engineered" appeared in over 60% of entries with near-zero correlation to domain competency. By anchoring textual tokens to formal ontology graphs (ESCO and South African OFO codes), our pipeline recovers true operational capacity.
Unstructured Text → Competency Projection
"I worked as a production supervisor for three years coordinating factory shifts, scheduling preventative machinery maintenance, enforcing ISO-9001 compliance, and training junior operators on PLC safety protocols."
High correlation on shift management & safety compliance
Substantial overlap in maintenance scheduling
Partial qualification via ISO-9001 standards
Where Lexical Ambiguity Challenges the Model
Documented Failure Modes
Genuine research maturity requires documenting failure modes with the same empirical scrutiny as benchmark successes:
"I managed several teams across production cycles ensuring daily targets were hit."
Diagnosis: The token classifier interprets 'managed' as executive administrative leadership rather than grounding the verb in factory shift operational constraints.
Mitigation: Conditioning entity weights on domain co-occurrence embeddings (e.g. 'production cycles' attenuates corporate leadership classification).
"NC(V) in Electrical Infrastructure Construction with mechanical fitting apprentice modules."
Diagnosis: Vocational programs integrate cross-disciplinary foundational modules. The classifier over-indexed on secondary mechanical tokens, downgrading the primary trade accreditation.
Mitigation: Enforcing strict National Qualification Framework (NQF) regulatory gazette hierarchies during graph projection.
"Candidate with 1 single observation: Score = 100% (1/1 items correct)."
Diagnosis: Arithmetic averages or maximum likelihood estimators report absolute certainty on a single correct response. Our Bayesian formulation maintains broad credible intervals rather than premature confidence.
Mitigation: Informative prior initialization with Minimum Information-Gain regularization.
@article{mahandana2026skillsintelligence,
author = {Mahandana, Daniel and Systems Lab Group},
title = {Skills Intelligence in TVET Ecosystems: Contextual Extraction and Standardized Ontologies},
journal = {The Sota Lab Working Papers},
year = {2026},
volume = {2},
number = {1},
url = {https://sota-lab.lovable.app/research/skills-intelligence}
}