Scope Summary
Congruent is seeking to support Generative AI Fabric development through its engineering team. The company’s engineers (“Vendor Personnel”) should have deep expertise in machine learning, large language models, advanced deep learning, data engineering, prompt optimization, and model fine-tuning. Familiarity with PyTorch or TensorFlow and experience deploying AI solutions on AWS, Azure, or GCP are essential. The team must also demonstrate a strong understanding of ethical AI practices, including bias mitigation, fairness, and regulatory compliance.
Project Summary
- Design and build production-grade Generative AI systems, including LLM, RAG, and agent-based applications.
- Develop robust data pipelines, evaluation frameworks, and ensure system reliability.
- Optimize model architecture and inference/training stacks for performance and cost efficiency.
- Manage large-scale OpenSearch clusters and build ingestion pipelines for high-quality indexing and retrieval.
- Create reusable solution patterns and reference implementations.
- Troubleshoot issues across the ML lifecycle: data preparation, training, fine-tuning, evaluation, and serving.
- Collaborate with product management, research, and engineering teams, present at events and conferences.
- Sprint-based goals to deliver software enhancements and bug fixes for 360E services.
- Ongoing support for the 360E customer base to resolve issues and determine implementation needs.
- Other support as determined in application release roadmap.
Technical Requirements
- BS/MS in Computer Science or related field; 10+ years of experience in distributed/cloud software development.
- Exceptional written and verbal communication skills.
- Demonstrated experience designing, implementing, and operating scalable AI/LLM systems in production.
- Hands-on expertise with LLM and Generative AI techniques, including parameter-efficient fine-tuning (e.g., LoRA/QLoRA), instruction tuning, and advanced prompt strategies.
- Practical experience with RAG architectures, vector search, and related tooling (OpenSearch, vector databases, PostgreSQL, Kafka).
- Proficiency in Python and shell scripting.
- Experience with LLM frameworks and LLMOps tools (LangChain, LlamaIndex, MLflow, KServe, Kubeflow, Triton).
- Proven ability to design data collection/annotation workflows and systematic evaluation for production quality.
- Strong publication record in top-tier venues is a plus.
Preferred Qualifications
- Advanced degree (MS/PhD) preferred.
- Expertise in RAG solution architecture and knowledge stores (OpenSearch + vector stores).
- Experience delivering Generative AI and agent solutions.
- Familiarity with multimodal modeling and modern computer vision.
- Experience with LLM serving technologies (e.g., DeepSpeed, FasterTransformer).
- Public cloud experience; strong understanding of IaaS/PaaS and competitive cloud capabilities.
- Experience with fine-tuning (PEFT) and multi-task serving techniques.
Deep understanding of Transformers, training methods, and optimizers; frameworks: PyTorch, JAX, or TensorFlow.