International Consultant - AI Data Analyst/Expert

DevJobs; Wien; Vor Ort; EUR 60.000 - 110.000; Vollzeit

Zusammenfassung

DevJobs sucht einen fachkundigen Consultant, um das SitRoom OSINT-Dataset der OSCE zu ingestieren, zu profilieren und in analyzing-ready daily feature tables zu überführen. Sie wenden ML-Algorithmen an, entwickeln Indikator-Generierungsmethoden und führen Back-Tests durch, um die Modelle zu validieren.

Zudem wird ein Prototyp entwickelt, der Analystenoutputs wie Berichte erzeugt. Sie erstellen Nutzer- und technische Dokumentation und setzen ein Forecasting-Modell für Early-Warning-Trends auf

Qualifikationen

• Fremdsprachen- und technische Dokumentation in Englisch können erforderlich sein.
• Datenanalyse, ML-Methoden an realen Datensätzen anwenden.
• Eigenständige Arbeitsweise und termingerechte Lieferung der Ergebnisse.

Aufgaben

• Daten ingestieren, profilieren und das SitRoom-Open-Source-News-Datenset verarbeiten.
• ML-Modelle auf dem Datensatz anwenden und verfeinern.
• Entwicklung einer Indikator-Generierungsmethode zur Erkennung auffälliger Termini, Akteure, Orte und Narrative.
• Historische Validierung und Back-Testing durchführen und Leistungsgrenzen dokumentieren.
• Prototypische Anwendung erstellen, die Analystenoutputs wie Berichte generieren kann.
• Nutzerdokumentation, technische Dokumentation und Übergabematerialien erstellen.
• Prognosemodell für Trendanalyse basierend auf SitRoom OSINT erstellen.

Kenntnisse

Analytische Schreibfähigkeiten

Englische Fachdokumentation

Eigenständige Arbeit

Datenanalyse & ML-Methoden

Ausbildung

First-level university degree in relevant field (z.B. Data Science, Statistik, Informatik)

Tools

Python

Jobbeschreibung

Deine Rolle im Team

• Under the overall supervision of the Chief, Operations Support Unit, the consultant will be tasked with the following duties:
• Ingest, profile and process the full SitRoom open-source news database made available by the OSCE, ensuring that the full archive can be transformed into analysis-ready daily feature tables suitable for early-warning analysis.
• Apply and refine Machine Learning algorithms on the full SitRoom open-source dataset.
• Develop and test the case-specific indicator generation method, using pre-defined periods and/or rolling prominence analysis to identify terms, actors, places, themes or narratives that become unusually prominent in a given case.
• Re-run historical validation and back-testing on selected OSCE-relevant cases, subject to data availability, and document the performance, limitations and transferability of the model.
• Develop a working prototype application that can generate analyst-facing outputs, including a summary report or exportable results.
• Prepare user documentation, technical documentation and handover materials enabling OSCE staff to understand, test and further develop the prototype.
• Develop a forecasting model for early warning trend analysis based on the SitRoom OSINT database.

Unser Angebot

• The initial duration of the assignment is until 31 March 2027.
• Further extensions will be subject to financial and operational requirements.
• Remuneration will be based on the selected consultant's/expert's qualifications, experience, the tasks and deliverables for this position and in accordance with the OSCE established rates.

Technologien und Skills

• Python

Unsere Erwartungen an dich:

Qualifikationen

• Ability to translate technical model outputs into clear, interpretable findings for non-technical users.
• Strong analytical writing skills and ability to prepare concise technical and policy-facing documentation in English.
• Ability to work independently and deliver outputs within tight timelines.

Erfahrung

• Demonstrated experience applying data analytics, statistical modelling and/or machine-learning methods to complex real-world datasets.
• Demonstrated experience with Python-based data analysis, including data processing, feature engineering, model development and visualization.
• Practical experience with Natural Language Processing (NLP) or text analytics, including one or more of the following: keyword-based feature extraction, topic modelling, sentiment analysis, named-entity recognition or document classification.
• Experience developing, testing or documenting analytical models using methods such as logistic regression, random forest, support vector machines, anomaly detection or related supervised/unsupervised techniques.
• Computer literacy and practical experience with Microsoft Office applications.

Ausbildung

• First-level university degree in business analytics, data science, statistics, computer science, economics, political science, international relations, security studies, conflict studies or another relevant field.