ML Engineer

Factagora; Republic (MO); Hybrid; USD 90,000 - 130,000; Full time

Job summary

Factagora is seeking an engineer to turn verifiable AI into reality by designing and improving ML pipelines behind its API endpoints, including Fact-checking, Evidence Finding, Causality Graph, and Timeseries extraction. Beyond training models, you'll help build our TKG system and evaluate reliability, tracking how facts change over time.

You’ll work on RAG retrieval and production deployment, applying ML/NLP to global sources.

Qualifications

• 2+ years ML/NLP experience or equivalent research and development.
• Production ML development with Python (PyTorch, HuggingFace).
• Experience applying an LLM or RAG system to a real service.
• NER, relation extraction, text classification in NLP.
• Systematic model evaluation and experiment management.
• Ability to read papers and implement ideas quickly.

Responsibilities

• Designing and improving ML pipelines for core API endpoints such as Fact-checking, Evidence Finding, and Deep Research.
• Developing unstructured text structuring models for FactBlock generation (NER, relation extraction, claim detection).
• Building the Temporal Knowledge Graph and tracking how facts change over time.
• Optimizing the RAG pipeline and improving retrieval accuracy and evidence reliability.
• Designing and improving an LLM-based multi-agent architecture (Agent Debate).
• Researching structured reasoning features like causality graphs and timeseries extraction.
• Building model evaluation frameworks and production deployment pipelines.
• Monitoring and automating knowledge graph quality from global sources.

Skills

2+ years ML/NLP

Python

PyTorch

HuggingFace

LLM or RAG systems

NER

Relation extraction

Claim detection

Model evaluation

Experiment management

Reading & applying papers

English communication

Tools

HuggingFace tooling

PyTorch tooling

Job description

An engineer who turns 'verifiable AI' into reality

You'll design and improve the ML pipelines behind Factagora's API endpoints, including Fact-checking, Evidence Finding, Causality Graph, and Timeseries extraction. Beyond training models, you'll help build our TKG (Temporal Knowledge Graph) system, which extracts knowledge from global news and research sources, evaluates reliability, and tracks how facts change over time.

Responsibilities

• Designing and improving ML pipelines for core API endpoints such as Fact-checking, Evidence Finding, and Deep Research
• Developing unstructured text structuring models for FactBlock generation (NER, Relation Extraction, Claim Detection, and more)
• Building the TKG (Temporal Knowledge Graph) and improving how it tracks facts changing over time
• Optimizing the RAG pipeline and improving retrieval accuracy and evidence reliability
• Designing and improving our LLM-based multi-agent architecture (Agent Debate)
• Researching and implementing structured reasoning features such as Causality Graph and Timeseries extraction
• Building model evaluation frameworks and operating production deployment pipelines
• Monitoring and automating knowledge graph quality from global news and research sources

Requirements

• 2+ years of ML/NLP work experience, or equivalent research and development experience
• Experience developing and deploying ML models in production with Python (PyTorch, HuggingFace, and similar)
• Experience applying an LLM or RAG system to a real service
• Modeling experience with NLP tasks such as text classification, named entity recognition, and relation extraction
• A systematic approach to model evaluation and experiment management
• Ability to read papers and quickly implement and apply them in practice

Preferred Qualifications

• Research or development experience with Knowledge Graphs or Graph Neural Networks
• Research experience in fact-checking, claim verification, or misinformation detection
• Experience implementing a multi-agent LLM system or a self-correcting pipeline
• Experience building large-scale news or document processing and information extraction pipelines
• Experience setting up MLOps environments and serving or monitoring models
• Experience operating a service for a global user base, or comfortable communicating in English
• Strong understanding of and interest in AI or deep-tech products

Who We're Looking For

• You're fundamentally curious about why AI-generated information gets things wrong
• You're comfortable turning ideas from papers into prototypes quickly and applying them in production
• You judge success by whether users can genuinely trust the result, not just model performance numbers
• You're comfortable breaking uncertain problems into small experiments and validating them quickly
• You genuinely connect with Factagora's mission of building a trustworthy AI and information ecosystem

Hiring Process

Application 1st Interview 2nd Interview Final Offer

Application Inquiries

If you run into any issues during the application process or have questions about this role, reach out to
[e-mail] and we'll get back to you as soon as possible.