Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Hi, I'm Benjamin Thomas, a Senior GenAI/LLM Engineer focused on building production-grade AI systems, RAG architectures, and scalable inference pipelines. I specialize in in-warehouse embeddings, vector search, and end-to-end orchestration of prompts, retrieved context, and model outputs to enable governed, auditable enterprise GenAI workflows. I enjoy bridging data, ML, and product teams to deliver measurable business impact, minimize hallucinations, and optimize cost and latency. My work emphasizes governance, observability, and a strong bias toward reproducibility and collaboration across cross-functional teams.…Hi, I'm Benjamin Thomas, a Senior GenAI/LLM Engineer focused on building production-grade AI systems, RAG architectures, and scalable inference pipelines. I specialize in in-warehouse embeddings, vector search, and end-to-end orchestration of prompts, retrieved context, and model outputs to enable governed, auditable enterprise GenAI workflows. I enjoy bridging data, ML, and product teams to deliver measurable business impact, minimize hallucinations, and optimize cost and latency. My work emphasizes governance, observability, and a strong bias toward reproducibility and collaboration across cross-functional teams.

Benjamin Thomas

AI Engineer, Data Scientist, Full Stack Developer, +4





Hi, I'm Benjamin Thomas, a Senior GenAI/LLM Engineer focused on building production-grade AI systems, RAG architectures, and scalable inference pipelines. I specialize in in-warehouse embeddings, vector search, and end-to-end orchestration of prompts, retrieved context, and model outputs to enable governed, auditable enterprise GenAI workflows. I enjoy bridging data, ML, and product teams to deliver measurable business impact, minimize hallucinations, and optimize cost and latency. My work emphasizes governance, observability, and a strong bias toward reproducibility and collaboration across cross-functional teams.…Hi, I'm Benjamin Thomas, a Senior GenAI/LLM Engineer focused on building production-grade AI systems, RAG architectures, and scalable inference pipelines. I specialize in in-warehouse embeddings, vector search, and end-to-end orchestration of prompts, retrieved context, and model outputs to enable governed, auditable enterprise GenAI workflows. I enjoy bridging data, ML, and product teams to deliver measurable business impact, minimize hallucinations, and optimize cost and latency. My work emphasizes governance, observability, and a strong bias toward reproducibility and collaboration across cross-functional teams.

Available to hire

Hi, I’m Benjamin Thomas, a Senior GenAI/LLM Engineer focused on building production-grade AI systems, RAG architectures, and scalable inference pipelines. I specialize in in-warehouse embeddings, vector search, and end-to-end orchestration of prompts, retrieved context, and model outputs to enable governed, auditable enterprise GenAI workflows.

I enjoy bridging data, ML, and product teams to deliver measurable business impact, minimize hallucinations, and optimize cost and latency. My work emphasizes governance, observability, and a strong bias toward reproducibility and collaboration across cross-functional teams.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Language

English

Fluent

Work Experience

Senior GenAI/LLM Engineer at Snowflake

January 1, 2022 - Present

Designed and deployed a Snowflake-native GenAI document intelligence platform using Retrieval-Augmented Generation (RAG), combining in-warehouse embeddings, vector similarity search, and LLM-based answer generation. Built Snowflake-native embedding pipelines using transformer-based sentence encoders (BERT-family, E5-style) via SnowparkMLUDFs, tuning token-aware chunk boundaries, overlap ratios, and embedding dimensionality to maximize Recall@K. Designed a governed LLM orchestration layer where prompts, retrieved context, and model outputs were materialized as Snowflake tables, enabling full lineage, reproducibility, and auditability for enterprise GenAI workflows. Implemented multi-stage retrieval pipelines combining ANN-based dense vector search, BM25-style lexical filters, and metadata-aware pruning to significantly reduce off-topic retrieval and downstream hallucinations. Developed a Snowflake-resident GenAI evaluation framework that joined retrieval metrics with LLM response scorin

Senior AI/ML Engineer - E-commerce Intelligence at Walmart Global Tech

January 1, 2020 - December 1, 2021

Built a recommendation engine for an E-Commerce platform using collaborative filtering and content-based filtering leveraging product metadata embeddings to personalize product listings and search rankings. Collaborated with Data Engineering to integrate Databricks and Airflow, streamlining data ingestion processes for ML workflows. Designed a visual outfit generation tool using HuggingFace Diffusion Models, which increased conversion rate by 18% by offering AI-generated, style-coherent outfits. Partnered with cross-functional teams to integrate ML solutions into client-facing e-commerce products, conducting A/B testing across product pages and checkout funnels to measure impact.

Machine Learning Engineer at Cognizant Technology Solutions

June 1, 2015 - December 1, 2019

Implemented GAN-based image augmentation for skin-cancer images to enhance datasets for deep learning models, improving detection accuracy by 20%. Pioneered a predictive analytics model (RNN/LSTM) to forecast patient admission rates, reducing emergency department congestion during peak times by 35%. Developed an autonomous decision-making system for drone navigation using reinforcement learning, simulating virtual environments to optimize route planning. Built an object detection system for drone-mounted cameras using YOLOv4, deployed in a conservation project for animal tracking and environmental monitoring. Automated model deployment processes using CI/CD pipelines (Jenkins and GitHub Actions), improving deployment speed and efficiency by 30% and ensuring seamless cloud integration.

SeniorGenAI/LLM Engineer at Snowflake

January 1, 2022 - Present

Designed and deployed a Snowflake-native GenAI document intelligence platform using Retrieval-Augmented Generation (RAG), combining in-warehouse embeddings, vector similarity search, and LLM-based answer generation. Built Snowflake-native embedding pipelines using transformer encoders (BERT-family, E5-style) executed via Snowpark UDFs, optimizing token-aware chunk boundaries, overlap ratios, and embedding dimensionality to maximize Recall@K. Designed a governed LLM orchestration layer where prompts, retrieved context, and model outputs were materialized as Snowflake tables, enabling full lineage, reproducibility, and auditability for enterprise GenAI workflows. Implemented multi-stage retrieval pipelines combining ANN-based dense vector search, BM25-style lexical filters, and metadata-aware pruning to reduce off-topic retrieval and downstream hallucinations. Developed a Snowflake-resident GenAI evaluation framework that joined retrieval metrics (Recall@K, MRR) with LLM response scoring