Looking for a job

Benjamin Jodoin

@benjaminjodoin

I build production RAG and LLM systems that improve enterprise support.

At Dust, I led the architecture and delivery of an enterprise RAG support assistant that reduced manual support workload by 45% and improved answer relevance by 35%.

I build end-to-end LLM systems, from document ingestion, chunking, embeddings, and vector retrieval through prompt construction, evaluation, and production deployment. My transformer fine-tuning improved response accuracy by 25%, while inference optimization reduced latency by 30%.

Previously at Cognizant and Scale AI, I built high-throughput semantic search, NLP pipelines, and Python data workflows. I've led six data scientists and ML engineers and delivered AI services on AWS using Docker and Kubernetes.

Experience

Work history, roles, and key accomplishments

GenAI & LLM Engineer

Dust

Sep 2025 - Jun 2026 (9 months)

Led architecture and delivery of an enterprise Generative AI support assistant using LLMs and RAG, reducing manual support workload by 45%. Designed end-to-end RAG pipelines and fine-tuned transformer models, improving accuracy and reducing latency.

Python PyTorch TensorFlow Hugging Face Transformers Lang Chain Faiss LLM RAG Embeddings Vector Search Docker Kubernetes AWS SQL PySpark

Team Lead & Machine Learning Engineer

Cognizant

Aug 2021 - Jul 2025 (3 years 11 months)

Designed and developed an LLM-based semantic search platform processing over 1 million queries per day with sub-second latency. Built transformer-based NLP pipelines and improved model performance by 20% through optimization.

Python BERT GPT PyTorch NLP Semantic Search Computer Vision CNN Docker AWS SQL PySpark Machine Learning Deep Learning

Python Developer - NLP / ML Focus

Scale AI

Jul 2018 - Jun 2021 (2 years 11 months)

Implemented NLP solutions for sentiment analysis, document classification, and text mining. Developed Python-based machine learning solutions for predictive analytics and automated decision support.