AI Developer

This is a fully remote position, open to applicants in Mexico.

📋 Description

• Train and refine LLMs through supervised fine-tuning (SFT).

• Collaborate with open-source models such as LLaMA, Mistral, Qwen, and other similar architectures.

• Develop LoRA/Q-LoRA pipelines to facilitate efficient fine-tuning.

• Implement and enhance data preprocessing workflows, encompassing tokenization and management of long-context data.

• Utilize and extend Hugging Face Transformers & Datasets for both training and inference tasks.

• Parse and manage structured and semi-structured data, including XML/XSD files.

• Create document parsing solutions for Office formats (python-docx, OpenXML).

• Deploy, operate, and sustain models in fully offline and air-gapped environments.

• Conduct model optimization and quantization (GGUF, GPTQ, AWQ, bitsandbytes).

• Develop and uphold inference systems using frameworks like vLLM, TGI, and Ollama.

• Optimize GPU utilization (CUDA, cuDNN, VRAM-aware batching).

• Maintain local CI/CD pipelines for ML models without relying on cloud services.

• Manage local model registries, version control, and artifacts.

• Create backend services in Python for ML training and inference workflows.

• Work with relational databases (Postgres/MySQL).

• Utilize Docker and Git for dependable development and deployment pipelines.

• Employ Azure DevOps for CI/CD processes (including local runners when appropriate).

⛳️ Requirements

• Extensive experience in Python for backend and machine learning development.

• Proficiency with ML frameworks such as PyTorch or TensorFlow, scikit-learn, and pandas.

• Strong understanding of Postgres or MySQL for data storage solutions.

• Experience with Docker, Git, and best practices in DevOps.

• Practical expertise in LLM training, fine-tuning, and optimization techniques.

• Familiarity with Hugging Face Transformers & Datasets.

• Knowledge of XML/XSD and tools for parsing Office documents.

• Experience in deploying models using vLLM, TGI, or Ollama.

• Understanding of quantization methods (GGUF/GPTQ/AWQ).

• Experience with GPU optimization and the CUDA ecosystem.

• Capability to develop solutions for offline, on-premises, and air-gapped environments.

• Nice to Have:

• Experience managing ML model registries in offline settings.

• Familiarity with AWS for hybrid deployment scenarios (not required).

• Background in secure environments, restricted networks, or enterprise compliance.

• Soft Skills:

• Strong sense of ownership and problem-solving skills.

• Ability to collaborate in distributed teams across various time zones.

• Effective communication skills when discussing complex technical subjects.

🏝️ Benefits

• Competitive salary and performance-based incentives.

• Opportunities for professional growth and development.

• Flexible working hours and remote work options.

• Collaborative and inclusive work environment.

• Access to cutting-edge tools and technologies.