AI Researcher, Engineer, & Architect
I bridge cutting‑edge AI research and production‑scale deployment. As Co‑Founder & Chief Data Scientist of AI Squared, I lead the creation of custom LLM families, robust guardrails, and agentic orchestration pipelines. I hold a PhD in Explainable AI and a TS/SCI clearance (Full‑Scope Polygraph), enabling me to deliver mission‑critical capabilities for the DoD and Intelligence Community.
Duties at AI Squared
I am the Chief Data Scientist and Cofounder at AI Squared. In this role, I help lead the development of the company's AI platform and its browser-based integration technologies. I've led R&D initiatives to advance generative AI models, including creating custom LLMs and contributing to the BeyondML project, which was released as open-source to the Linux Foundation Data and AI.
Duties at Capitol Technology University
At Capitol Technology University, I serve as Adjunct Faculty, mentoring PhD students in AI research and teaching undergraduate courses like CS-150. My research focus includes generative AI, transformer models, and their applications to fields like agriculture. I also serve as an external examiner for dissertation defenses and support the development of doctoral research in AI.
Key Projects & Impact
Bolt LLM Family
Architected and delivered a suite of domain‑tuned LLMs with built‑in guardrails, on‑premise efficiency, and a modular API layer that powers AI Squared’s UNIFI platform.
BeyondML (Linux Foundation)
Open‑sourced a novel multi‑task neural architecture that isolates subnetworks for independent tasks—adopted by the Linux Foundation Data & AI community.
Agentic Orchestration Engine
Built a closed‑loop, tool‑use environment enabling LLMs to execute multi‑step workflows with dynamic skill extensions and secure system‑prompt configuration.
AI Squared UNIFI Platform
Delivered production‑grade APIs for embeddings, translation, chat, and a guardrails engine; supports on‑premise GPU clusters and HPC environments.
Technical Skills
- AI/ML: LLMs, XAI, Agentic Frameworks, PEFT/LoRA, RLHF, BERT, RoBERTa, LSI, Topic Modeling
- LLM Engineering: Quantization (GGUF, bitsandbytes), vLLM, TGI, Vector DBs, RAG, Advanced Chunking
- Frameworks: PyTorch, HuggingFace, FastAPI, Dask, Scikit‑learn, Pandas, NumPy
- Infra / MLOps: Docker, MLflow, W&B, GitHub Actions, DeepSpeed, Slurm, RabbitMQ, HPC, GPU acceleration
- Languages: Python, JavaScript, C, LaTeX
- Specializations: Explainability, Guardrails, Model Routing, Production RAG, LLM‑as‑Judge
Open‑Source Contributions
Over the years, I've contributed to and created multiple open-source projects. Some highlights include:
Education
I hold a PhD in Technology from Capitol Technology University, with a focus on Explainable AI. My dissertation explored linear regression feature engineering in classification trees. Additionally, I have a Master of Science in Business Analytics from the University of Maryland, and a Bachelor of Science in Mathematics from the same institution.
Selection of Talks and Publications
- Jacob Renn and Ian Sotnek (2023, June 26-29). "D-Lite: Integrating a Lightweight ChatGPT-Like Model Based on Dolly into Organizational Workflows" [Conference session]. Data+AI Summit 2023, San Francisco, CA, United States.
- Jacob Renn and John Daly (2022, October 17-19). "Using Artificial Intelligence to Enrich Experiences in the Metaverse" [Conference session]. O3DCon 2022, Austin, TX, United States.
- Jacob Renn and John Daly (2022, October 17-19). "Multitask Machine Learning and its Applications to the Metaverse" [Conference session]. O3DCon 2022, Austin, TX, United States.
- Ian Sotnek and Jacob Renn (2022, June 27-30). "Unifying Data Science and Business: Artificial Intelligence Augmentation and Integration into Production Business Applications" [Conference session]. Data+AI Summit 2022, San Francisco, CA, United States.
- Jacob Renn and Ian Sotnek and Benjamin Harvey and Brian Caffo (2022). "The Multiple Subnetwork Hypothesis: Enabling Multidomain Learning by Isolation Task-Specific Subnetworks in Feedforward Neural Networks" [arXiv preprint].
- Jacob Renn (2022). "Linear Regression Feature Engineering in Classification Tree Learning" [Dissertation]. Capitol Technology University, Laurel, MD, United States.
