LLM Research Intern

About ProCogia: 

The core of our culture is maintaining a high level of cultural equality throughout the company. Our diversity and differences allow us to create innovative and effective data solutions for our clients. 

Our Core Values: Trust, Growth, Innovation, Excellence, and Ownership

Responsibilities 

• Assess client-specific data assets and determine the appropriate adaptation strategy — continued pretraining, supervised fine-tuning, or a combination — based on the domain, data volume, and use case requirements 

• Curate, clean, structure, and prepare domain-specific datasets from raw client data for use in model training pipelines 

• Fine-tune large language models in the 70B–100B+ parameter range using techniques such as LoRA, QLoRA, and multi-adapter patterns 

• Perform continued pretraining on open-weight models (Qwen, Llama, and related ecosystems) to embed domain knowledge directly into model weights 

• Manage distributed training workflows across multi-node GPU clusters 

• Design and execute evaluation frameworks to validate domain adaptation quality, factual grounding, and model behavior 

• Support RAG system development where applicable, including vector database integration, chunking strategies, and reranking pipelines 

• Contribute to inference optimization and deployment pipeline integration 

Required Qualifications 

• Currently enrolled in or recently completed a Bachelor's, Master's, or PhD program in Computer Science, Machine Learning, or a related field 

• Demonstrated hands-on experience fine-tuning large language models, supported by concrete project work, research, or open-source contributions 

• Experience with frontier-scale models (100B+ parameters) or distributed training across multi-node GPU clusters 

• Familiarity with parameter-efficient fine-tuning methods (LoRA, QLoRA) and open-weight model architectures 

• Experience with data curation and preparation workflows for LLM training, including cleaning, formatting, deduplication, and quality filtering 

• Proficiency in Python-based ML frameworks such as PyTorch, HuggingFace Transformers, DeepSpeed, or FSDP 

• Understanding of training compute, memory constraints, and inference trade-offs at scale 

Nice to Have 

• Familiarity with RAG architectures or production inference serving frameworks (vLLM, TGI, TensorRT-LLM) 

• Experience in low-resource or multilingual NLP settings 
• Relevant publications, open-source contributions, or documented projects involving LLM training 

ProCogia is proud to be an equal-opportunity employer. We are committed to creating a diverse and inclusive workspace. All qualified applicants will receive consideration for employment without regard to race, national origin, gender, gender identity, sexual orientation, protected veteran status, disability, age, or other legally protected status.