About ComplyForge Technologies

ComplyForge builds AssuraQMS, an AI-powered electronic Quality Management System for medical device, IVD, and health-technology companies. The platform integrates Document Management, Design Controls, Risk Management, CAPA, Nonconformance, Supplier Controls, Training, Audits, Complaints, and Post-Market Surveillance. AssuraQMS is designed around quality-system processes aligned with ISO 13485:2016, FDA QMSR, 21 CFR Part 11, and ISO 14971.

We are looking for an AI/ML Engineer Intern who is eager to work on real-world AI problems where accuracy, evidence, traceability, and reliability are paramount.

About the Role: AI Engineering Intern

This is a full-time internship for 6 months, with a potential pathway to a full-time role based on performance and business requirements.

What You’ll Work On

RAG & Regulatory Intelligence Build and improve retrieval pipelines over regulatory, technical, and customer-controlled information. This may involve:

• Document ingestion and preprocessing
• Chunking and embeddings
• Vector and hybrid search
• Metadata filtering
• Ranking and reranking
• Context construction
• Retrieval-quality evaluation

Sources may include FDA information, regulatory databases, medical-device documentation, quality procedures, and controlled documents.

510(k) Predicate Finder Contribute to our 510(k) Predicate Finder for predicate-device research and substantial-equivalence analysis. This may involve:

• Structured data extraction
• Device-attribute normalisation
• Similarity and ranking
• Evidence retrieval and filtering
• Citation generation
• Structured device comparisons

Citation Grounding & Hallucination Control In regulated workflows, unsupported AI output is unacceptable. You’ll help develop systems that:

• Connect AI outputs to supporting evidence
• Validate citation relevance
• Detect unsupported claims
• Identify fabricated references
• Test failure and abstention behaviour
• Improve retrieval and generation quality

AI-Assisted Quality Workflows Help develop AI capabilities for:

• SOP and controlled-document assistance
• CAPA investigation support
• Design-control traceability
• Regulatory and quality intelligence

The goal is to provide useful AI assistance while maintaining traceability, reviewability, and human control.

Model & Prompt Lifecycle Help make AI-assisted workflows reproducible and inspectable by tracking:

• Models
• Prompts
• Retrieval Configuration
• Index Versions
• Parameters
• Evaluation Results
• Source Evidence

Requirements

Essential

• Pursuing or recently completed a degree in Computer Science, AI, ML, Data Science, Electronics, or a related field.
• Strong Python fundamentals.
• Ability to work with an existing codebase.
• Practical experience with LLM APIs.
• Understanding of prompt design and structured model interaction.
• Experience building at least one working RAG/retrieval-based application.
• Understanding of embeddings and vector search.
• Familiarity with Git.
• Strong analytical and problem-solving skills.
• Curiosity about why an AI system produces a particular result.

Your experience can come from a university project, personal project, hackathon, or research project.

Good to Have

• Experience with LangChain, LlamaIndex, or similar frameworks.
• Experience with pgvector, FAISS, Qdrant, or other vector databases.
• Familiarity with hybrid search and reranking.
• Experience with LLM evaluation frameworks.
• Skills in document parsing and OCR, including PDF/table extraction.
• Experience with structured outputs and tool/function calling.
• Familiarity with FastAPI or backend development.
• Knowledge of PostgreSQL.
• Experience with TypeScript / React.
• Familiarity with AWS, Azure, or GCP.

What You’ll Learn

You’ll gain practical experience in:

• RAG
• AI Evaluation
• Evidence-Grounded AI
• Hallucination Control
• Document Intelligence
• Human-in-the-Loop AI
• AI Lifecycle Management

You’ll also develop working knowledge of ISO 13485, FDA QMSR, 21 CFR Part 11, ISO 14971, and IEC 62304. Most importantly, you’ll learn how AI engineering changes when the output needs to be defensible, traceable, and reliable.

Why ComplyForge?

AI for regulated industries requires more than impressive demos; it requires evidence, traceability, controlled change, human review, and measurable performance. At ComplyForge, you’ll work on a real AI-powered product solving real quality and regulatory problems for the medical-device industry. You’ll be working at the intersection of AI + Medical Devices + Regulatory Technology + Quality Systems. This is an opportunity to build a specialised skill set early in your career.

How to Apply

Please send us:

• Your CV
• A link to something you've built (GitHub, notebook, project page, technical write-up, or equivalent).
• A short note (max 200 words) describing one AI system you have built or studied.

Previous medical-device or regulated-industry experience is not required. We’re interested in what you’ve built, how you think, and how deeply you understand the systems you work with.