Computational Linguist

Descripción del empleo

At

Indeep AI

, we are building the DPDF (Dynamic Portable Document Format) ecosystem. This project moves beyond standard text processing to create a new paradigm of digital text that evaluates documents for pragmatic honesty, logical fallacies, and cognitive biases. We are redefining how humans interact with written documents by introducing conversational agents grounded in Grice’s Cooperative Principle and Theory of Mind.

About the Role

We are seeking a Computational Linguist to join our team at Indeep Artificial Intelligence (IAI) who is as comfortable with formal semantics as they are with high-performance software engineering. You will be responsible for the "Brain" of the DPDF format: the engine that ingests raw text and outputs a structured, multi-layered pragmatic analysis. You are not just a user of Large Language Models (LLMs); you are an architect who can programmatically constrain, evaluate, and chain them to perform complex linguistic "surgeries" on documents.

We also need someone with a strong startup attitude. In practice, this means:

• A builder mindset: You don’t just “request” features; you build them. You thrive in an environment where the codebase is evolving and the problem space is undefined.
• Versatility: You are willing to wear the hat of a researcher one hour and a DevOps engineer the next, ensuring that your linguistic pipelines are performant, scalable, and secure.
• You are precision-obsessed: In a project focused on honesty and clarity, your code and prompts must be the gold standard of precision.

This role is for someone energised by wearing multiple hats and making an impact without needing a detailed playbook. If that sounds like you, we’d be excited to hear from you.

Key Responsibilities

Corpus Building & Annotation

• Design, build, and curate annotated corpora of domain-specific documents, including insurance contracts, legal texts, and grant applications.
• Define and maintain annotation schemas for pragmatic, semantic, and structural features, ensuring inter-annotator reliability.
• Oversee and quality-control annotation workflows, including work performed by external annotators or domain experts.

Pragmatic Analysis Engine

• Develop and implement the core logic that classifies text according to Gricean Maxims (Quality, Quantity, Relation, Manner) and identifies Epistemic Traps.
• Design and implement software functions for granular document evaluation, automated Pragmatic Tagging, and real-time text editing and simplification.
• Build proprietary linguistic metrics.

Advanced Prompt Engineering & LLM Orchestration

• Architect complex LLM orchestration pipelines, including Chain-of-Thought (CoT) & ReAct, programmatic prompting, few-shot, and meta-prompting.
• Manage LLM hallucinations through RAG (Retrieval-Augmented Generation) and semantic validation layers.

Linguistic Analysis Pipeline Design

• Design and refine the linguistic analysis pipelines powering our TRIC scoring system (Truthfulness, Relevance, Informativeness, Clarity) and its Grice’s maxim-based feedback engine.
• Translate pragmatic and discourse-level linguistic theory into concrete, implementable feature specifications.
• Maintain and evolve linguistic rule sets, ontologies, and concept taxonomies across document types and target languages (initially Italian, Spanish, and English).

Evaluation Frameworks & LLM Judging

• Develop rigorous evaluation frameworks to assess the quality, accuracy, and consistency of LLM-generated outputs.
• Design and implement LLM-as-judge pipelines, including prompt design, evaluation rubrics, and inter-rater calibration against human expert baselines.
• Conduct systematic error analysis to identify failure modes in model outputs and translate findings into actionable improvement cycles.

Cross-functional Collaboration

• Work closely with the product team to inform model fine-tuning, feature engineering, and retrieval pipeline design with linguistic expertise.
• Collaborate with the product and compliance teams to ensure linguistic outputs meet regulatory and accessibility requirements.

Qualifications Required

• PhD or MSc in Computational Linguistics, NLP, or Computer Science with a strong NLP focus. A solid STEM foundation is non-negotiable.
• Expert-level proficiency in Python. Hands-on experience with NLP libraries (HuggingFace, spaCy, NLTK) and LLM frameworks (LangChain, LlamaIndex, or equivalent).
• Deep understanding of pragmatics and semantics. Ability to translate abstract principles, such as the Cooperative Principle, into code-based constraints.
• Proven experience managing LLM outputs through RAG pipelines and semantic validation layers.
• Demonstrated ability to design and implement complex prompting strategies for LLMs, including evaluation and judging pipelines.
• Ability to design validation tests for linguistic models, ensuring high inter-rater reliability between AI and human expert baselines.
• Hands-on experience building and curating annotated corpora, including annotation schema design and quality control.
• C2-level proficiency in English.

Qualifications Preferred

• Experience with readability metrics and automated text quality assessment (e.g., Flesch-Kincaid, Gulpease).
• Background in legal, insurance, or scientific/grant document processing.
• Experience designing inter-annotator agreement studies and applying relevant metrics.
• Working proficiency in Italian and/or Spanish.
• Familiarity with EU AI Act requirements or regulatory NLP applications.

What We Offer

• €50,000–€80,000 yearly gross salary depending on demonstrable experience and depth of expertise.
• Employee Stock Ownership Plans (ESOPs).
• A role at the intersection of cutting-edge NLP research and real-world product impact.
• Early-stage team member at a high-potential EIC-funded startup.

How to Apply

Submit your CV and a brief cover letter explaining your linguistic expertise and why you’re excited about this role through LinkedIn. Applications will be reviewed on a rolling basis, so apply now to join the innovative team at Indeep AI!

Información extra

Status

Activa

Estudios requeridos

E.S.O

Localización

Barcelona

Tipo de contrato

Trabajo estudiantes

Publicado el

05-04-2026

Carnet de conducir

No

Vehículo

No

Carta de motivación

No

Idiomas

Español