NLP/Linguistics Software Engineer

• Implement and Maintain: Write high-quality, maintainable code to support the analytics platform and its record matching components.

• Bridge Theory and Practice: Take theoretical ideas from linguistics and data science and implement them as practical software features.

• Support Search Internals: Help optimize and maintain search engine components, including Elasticsearch data modeling and performance tuning.

• Collaborate and Learn: Participate in agile sprint planning and work daily with senior partners to translate project requirements into technical solutions.

• Build Scalable Systems: Assist in designing and shipping robust APIs and scalable architectures that integrate into our AI-native platform.

Required:

• 2–4 years of professional software engineering experience (including high-impact internships or projects).

• Proficiency in Java (our core analytics language) or Python (for AI/ML integrations).

• Problem Solver: Ability to work across teams and make steady progress in ambiguous problem spaces.

• Educational Foundation: Bachelor's degree in Computer Science, Linguistics, or a related technical field.

Preferred (Nice to Have):

• Foundation in Data Science: Experience with data quality evaluation, data annotation, or guideline design, preferably for linguistics.

• Familiarity with Elasticsearch internals or other search/retrieval-based systems.

• Exposure to computational linguistics or natural language processing (NLP).

• Interest in Kubernetes and cloud-native architectures.
• Month 1–2: Ramp up on the analytics stack and record matching architecture; ship your first initial changes to production.

• Month 3–4: Take ownership of a specific component or pipeline improvement with guidance, including full testing and documentation.

• Month 5–6: Deliver a measurable improvement to record matching quality or pipeline reliability and contribute to team design discussions.

The record matching functionality is where Babel Street’s signals become usable intelligence. Do you care about provenance, explainability, and trust? When a match decision affects whether someone is onboarded or investigated, "the model said so" is not good enough. You will help build systems where every match is defensible, auditable, and tunable — a rare luxury in modern ML-heavy stacks. Do you speak multiple languages? Since our platform processes data from around the globe, your linguistic insights can directly inform how we build and polish the NLP and computational linguistics components that make our record matching world-class.