We are seeking an experienced Machine Learning Engineer (LLMs & NLP) to build document intelligence and reasoning pipelines. This role focuses on extracting structured, verifiable data from text-heavy and semi-structured documents such as specifications, invoices, reports, and technical documentation.
The emphasis is on high-precision extraction, document version awareness, and audit-ready outputs, not conversational chat systems.
Key Responsibilities
- Design and implement LLM-powered pipelines for document understanding and structured data extraction
- Build and maintain RAG workflows over large collections of versioned project documents
- Extract text, tables, metadata, and domain-specific fields with page-level traceability
- Implement certainty scoring, conflict detection, and human-in-the-loop validation logic
- Ensure outputs are deterministic, structured, and linked to exact document versions and project stages
- Collaborate with Data Engineers and Vision ML Engineers
- Optimize prompts, retrieval strategies, chunking, and ranking for high accuracy
- Monitor and continuously improve extraction quality in production
Required Qualifications for Medior
- Strong experience with LLMs, NLP, or document intelligence systems
- Proficiency in Python and modern ML tooling
- Hands-on experience with RAG architectures and vector search
- Solid understanding of structured outputs, schema enforcement, and validation workflows
- Experience working with complex, long-form, or regulated documents
- Strong analytical thinking and attention to detail
- English level B2 or higher
Additional Requirements for Senior / Lead Profiles
- Experience with LLM fine-tuning (SFT, LoRA, adapters) and model evaluation
- Experience deploying LLM-powered systems in production at scale
- Strong understanding of LLM infrastructure, cost optimization, and monitoring
- Experience with cloud platforms (AWS and/or Azure), including managed LLM services
- Experience designing API-based AI services (e.g. FastAPI)
- Ability to define AI architecture, mentor engineers, and lead technical decisions
What We Offer
- Opportunity to work on complex, real-world data engineering challenges with direct impact on the product
- Modern and evolving tech stack with room to influence architectural and technical decisions
- Continuous education and professional development, including access to courses, training programs, conferences, and learning resources
- Support for developing both technical and soft skills, with mentorship and knowledge-sharing within the team
- Collaborative and engineering-driven work culture
- Hybrid work model with flexible work-from-home arrangements
- Compensation based on experience and expertise
