• Design post-training pipelines (instruction tuning, RLHF, evaluation loops, etc.)

• Implement and optimize training workflows using open-source frameworks

• Experiment with model architecture improvements and hyperparameter optimization

⸻

AI Agent & System Integration

• Build and improve AI agents capable of executing multi-step workflows

• Integrate models into production environments and product features

• Improve model reliability, accuracy, and robustness for high-stakes applications

• Develop evaluation and testing frameworks for model performance and edge cases

⸻

Data & Experimentation

• Work with proprietary domain datasets to improve model specialization

• Design training datasets, labeling pipelines, and data quality frameworks

• Conduct ablation studies and performance benchmarking

⸻

Cross-Functional Collaboration

• Partner with product teams to design AI-first product experiences

• Provide technical feasibility insights for roadmap and feature planning

• Collaborate with global engineering teams on implementation and scaling

⸻

Required Qualifications

• 1–4 years of hands-on experience training or fine-tuning ML/AI models

• Experience working with LLM or multimodal model training pipelines

• Strong Python skills and familiarity with deep learning frameworks such as:

• PyTorch

• Hugging Face ecosystem

• Open-source training frameworks

• Experience deploying ML models into production systems

• Understanding of evaluation methodologies and model performance tradeoffs

⸻

Preferred Qualifications

• Experience with post-training techniques such as:

• Instruction tuning

• RLHF / preference optimization

• Model distillation

• Experience building AI agent or workflow automation systems

• Familiarity with distributed training and GPU optimization

• Experience working with domain-specific data (finance, compliance, enterprise workflows, etc.)

• Background working in early-stage startups or research labs

⸻

Nice to Have (Bonus Skills)

• Exposure to pre-training or large-scale model training environments

• Experience with retrieval-augmented generation (RAG)

• Experience designing evaluation benchmarks for enterprise AI applications

• Experience working with multi-agent systems or tool-use models