Transflo is seeking an experienced Machine Learning Engineer specializing in AWS Bedrock, MLflow, and advanced prompt engineering methodologies. The role involves leading the development of state-of-the-art MultiModal Document Identification and Extraction solutions, designing and fine-tuning foundation models, and implementing Generative AI strategies for efficient multimodal document processing.
Responsibilities:
- Design, develop, and deploy scalable machine learning models using AWS Bedrock and SageMaker
- Implement and optimize multimodal machine learning pipelines for document identification and extraction
- Develop and refine advanced prompt engineering strategies, including hierarchical prompting, context-aware prompts, and multi-turn dialogue techniques, to enhance the performance of foundation models
- Manage the end-to-end ML lifecycle, including experiment tracking, model versioning, and deployment using MLflow
- Ensure robust MLOps practices, including CI/CD pipelines, model monitoring, and automated retraining workflows
- Optimize model inference performance and cost-effectiveness using AWS Elastic Inference and SageMaker optimization techniques
- Integrate AWS Textract and Rekognition for enhanced OCR and image processing within ML workflows
- Collaborate with cross-functional teams, including data scientists, cloud engineers, and business stakeholders, to align AI models with business objectives
- Monitor, debug, and enhance machine learning workflows for improved reliability and efficiency
- Stay updated on the latest advancements in AI, multimodal machine learning, and AWS technologies, and apply them to real-world problems
Requirements:
- Extensive experience with AWS Bedrock for deploying and fine-tuning foundation models (FMs) for multimodal applications
- Proficiency in Amazon SageMaker for training complex ML models, hyperparameter tuning, and scalable deployment
- Hands-on experience with MLflow in AWS for experiment tracking, model versioning, and end-to-end ML lifecycle management
- Experience with AWS Lambda, API Gateway, and Step Functions for building serverless AI pipelines
- Familiarity with AWS Textract and Amazon Rekognition for document extraction and image recognition tasks
- Proficient in the utilization of Textual Models for Image Classification or other Open Source Image Classification tools
- Proficiency in AWS Deep Learning AMIs for rapid ML environment setup
- Experience with Amazon Elastic Inference for cost-effective inference acceleration
- Image-Text Alignment Prompts – Creating prompts that effectively link textual and visual data for accurate information extraction
- Hierarchical Prompting – Designing prompts for complex document structures with nested elements
- Context-Aware Prompting – Developing prompts that adapt to the semantic context of documents
- Visual Layout-Aware Prompting – Crafting prompts that leverage document layout information for precise entity recognition
- Few-shot and Zero-shot Prompting – Utilizing examples to improve multimodal model performance with minimal labeled data
- Multi-turn Dialogue Prompting – Implementing iterative prompts for complex document extraction scenarios
- Cross-Attention Prompts – Optimizing attention mechanisms for aligning visual and textual features
- Results oriented
- Independently reliable; performs tasks without close supervision
- Persistent Learner showing a desire to be on the edge of new AI methodologies as it may relate to current business opportunities
- Organized; detail-oriented, methodical and consistently demonstrates ability to successfully and timely complete assignments
- Follows-Up; consistently performs this in a positive, proactive manner
- Logical problem-solving skills
- Quality conscious; consistently demonstrates commitment to customers & quality
- Demonstrates timeliness & urgency
- Team work; individual contributor that works well with other team members and consistently promotes a strong team environment work ethic
- Goal setting; sets/achieves goals and consistently demonstrates a willingness/dedication to process improvement
- Responsible; takes responsibility for personal actions and consistently demonstrates a willingness to accept greater project responsibilities
- Professionally candid communications
- Focused on key success factors
- Professional attitude; consistently demonstrates ability to accept criticism and manage the conversation appropriately
- Street smart; can apply knowledge and life experiences in business
- Positive attitude
- Flexible & adaptable
- Resourceful