Open to opportunities

Sam User

@samuser29

Message

AI/ML developer focused on LLM agents, RAG, and computer vision—turning research into production-ready systems.

United States

Message

What I'm looking for

I’m looking for a team where I can build AI systems end-to-end—LLM agents with RAG, strong evaluation, and production reliability—while collaborating on applied research that ships.

I’m an AI/ML developer building conversational and agentic systems that combine strong research foundations with practical production engineering. My focus is end-to-end delivery: from data curation and model training to reliable orchestration and deployment.

As an AI Developer Intern at Hirello, I developed a conversational voice agent using Twilio and ElevenLabs, integrated with a platform RAG-powered Interview Agent and orchestration layer to simulate real interview conditions. I also curated structured fine-tuning datasets from interaction logs, retrained with Vertex AI/Gemini, and improved interview-query response accuracy by 18%.

At Purdue, I work as a Research Assistant—benchmarking sequence-fusion architectures on large GPS datasets, building robust multi-label pipelines (including leakage detection and correction), and optimizing GPU training by pre-computing cached ResNet-18 satellite features across many runs. I’ve also been accepted to present at ASABE AIM 2026 for my field operation analysis framework.

Previously, as a Machine Learning Intern at IQPaths, I built an end-to-end computer vision pipeline for early-stage soybean disease detection (93% recall), developed a collaborative-filtering recommendation system for nutritionally balanced 7-day meal plans, and authored hands-on ML labs that improved course completion by 20%. My projects include GRPO for LLM fine-tuning and a cost-constrained LangGraph agent runtime with budget enforcement and checkpoint replay.

Experience

Work history, roles, and key accomplishments

Current

AI Developer Intern

Current

Hirello LLC

Jun 2026 - Present (1 month)

Developing a conversational voice agent for real-time mock interviews using Twilio and ElevenLabs, integrated with a RAG-powered interview agent and orchestration layer. Curated and fine-tuned a structured Q&A dataset on Google Vertex AI with Gemini (18% improved accuracy) and built an Opportunities Agent for lead detection and recruiter outreach automation.

Twilio ElevenLabs Retrieval Augmented Generation (RAG)Fine Tuning Vertex AI Gemini Web Scraping Pinecone Orchestration

Current

Research Assistant

Current

Purdue University

Dec 2025 - Present (7 months)

Benchmarks and builds machine learning models using GPS tracks and satellite imagery, including sequence-fusion architectures and multi-label classification pipelines. Pre-computes and caches ResNet-18 satellite image features to reduce redundant GPU compute across multiple training runs and is accepted to present research at ASABE AIM 2026.

BiLSTM Transformers Cross Attention Weighted BCE GPU Training Bi LSTM

Machine Learning Intern

IQPaths Technologies Pvt. Ltd.

Oct 2024 - Dec 2024 (2 months)

Built an end-to-end computer vision pipeline using ResNet-50 for early-stage soybean disease detection, achieving 93% recall and reducing false negatives in field testing. Developed a recommendation system to generate nutritionally balanced 7-day meal plans for 5 hostels (60% fewer meal-plan revisions) and authored ML coding labs to improve learner engagement and course completion by 20%.

Computer Vision Disease Detection Recommendation Systems Collaborative Filtering data preprocessing Model Evaluation Educational Content (ML Coding Labs)