AI Solutions Engineer - Nashville, TN (Hybrid)

Our client is looking AI Solutions Engineer Real-Time Image Processing & Generative AI for Long term project Nashville TN below is the detailed requirements.

Job Title : AI Solutions Engineer Real-Time Image Processing & Generative AI

Location : Nashville TN

Duration : Long term

Job description:

We are seeking a skilled AI Solutions Engineer with expertise in real-time image processing, Generative AI frameworks (like AutoGen), and strong understanding of retrieval techniques, neural networks, and ML fundamentals. You will contribute to building scalable, modular, AI-driven solutions for image discrepancy detection and reporting.

Key Responsibilities

Real-Time Image Processing & Discrepancy Detection

• Implement real-time workflows where image uploads to AWS S3 trigger Lambda functions for immediate discrepancy detection.
• Prepare for high-throughput scenarios by designing scalable infrastructure with SQS and load-balanced Lambda triggers.
• Manage image transmission by converting images to byte format with minimal preprocessing to preserve context integrity.

GenAI & Agentic Frameworks

• Utilize agentic frameworks (e.g., AutoGen) for modular task separation including object detection, rule retrieval, and reporting.
• Justify design choices for modularity, scalability, maintainability, and debugging over monolithic LLM prompts.

Retrieval-Augmented Generation (RAG)

• Implement chunking strategies: fixed-word with overlap, semantic chunking, and rule-based segmentation.
• Integrate semantic and hybrid retrieval approaches including Amazon Kendra, cosine similarity, and metadata filtering.
• Build graph-based RAG pipelines by extracting data from PDFs/images and transforming it into structured knowledge graphs.

LLM Knowledge & Application

• Demonstrate understanding of Transformer architecture, self-attention, and token prediction mechanisms.
• Tune model behavior using temperature settings and explain its mathematical impact on output variability.
• Optimize prompt engineering and retrieval strategies for LLM use cases.

Neural Networks & ML Techniques

• Address vanishing/exploding gradients via ReLU, batch normalization, gradient clipping, and smart initialization (Xavier/He).
• Apply optimization algorithms like SGD, Adam, RMSProp for model convergence.
• Employ ensembling techniques like bagging and boosting to tackle overfitting and underfitting.

Data Analysis & Preprocessing

• Detect and manage outliers through Z-scores, IQR, and transformation techniques.
• Assess feature dependencies using correlation matrices, chi-squared tests, and statistical hypothesis testing.
• Interpret kurtosis values to evaluate tail distributions in datasets (mesokurtic, leptokurtic, platykurtic).

Required Skills

• AWS (S3, Lambda, SQS)
• Python (Byte handling, Model endpoints)
• Experience with AutoGen or similar agentic AI frameworks
• LLM application with retrieval techniques (RAG, Kendra, vector DBs)
• Strong fundamentals in ML, neural networks, and optimization
• Data preprocessing and statistical analysis