Ilia Ilmer, PhD
I am a Machine Learning Engineer working on NLP problems in Legal Tech space. My professional interests span everything from software development, machine learning, AI, to physics and mathematics to history and languages.
Experience
Machine Learning Engineer II, Clio, June 2024 - Present
Worked on applying machine learning and AI to solve business problems for law firms of all sizes. Deployed models enhancing new client intake experience, improved Clio Duo chatbot performance by utilizing latest prompt engineering techniques.
- Developed intelligent client intake chat agent (based on GPT-4.1) for Clio Grow customers
- Led development and deployment an XGBoost-based lead classification model for Clio Grow customers, achieving 92% accuracy on spam detection and 73% precision across three fit categories
- Developed marketing campaign email generating capability for Clio Grow customers
- Standardized and expanded test coverage for Clio’s intelligent assistant, Duo, doubling pass rates
Senior Machine Learning Engineer, Bill.com, February 2023 - June 2024
Improved model performance when processing users’ financial documents, increasing accuracy and correctness rates across multiple fields thus enhancing user experience and driving company revenue.
- Utilized Large Language Models for document processing (understanding, summarization)
- Decreased duplicate and false document matches by 50%, driving revenue up by over $3 million monthly
- Developed an improved field extraction model, raising accuracy from 35% to over 52% per field without latency decrease
- Mentored junior new hire colleagues and interns
Technical Experience
smol-autograd- An automatic gradient library written in C, with a minimal Adam implementation
tasklist- a personal utility for task management, written in Rust
indexer- a directory indexer, written in C learning about red-black trees
tinyflow- a flow matching implementation written using Python and Tinygrad (no PyTorch)
nanoGPT- a re-write of A. Karpathy’s GPT Tutorial in Python with Tinygrad (no PyTorch)
attention- implementing various attention algorithms from scratch to train on Shakespeare (based on papers only)
bpe- a learning project implementing BPE vocab building in C
chip8- a partial CHIP-8 CPU Emulator written in Rust
Programming Languages
- Python:
- Advanced proficiency, primary language of work
- Julia:
- Intermediate to advanced proficiency, primary language used in PhD work
- C/C++:
- Intermediate proficiency, language I like to use for personal projects when learning to build things from scratch
- Rust:
- Programming language I am currently learning (by building small projects).
Publications
- Google Scholar
- A list of my publications is available on Google Scholar
Education
- 2018-2023
PhD, Computer Science; CUNY Graduate Center (New York, NY)
Thesis title: Accelerating Parameter Identifiability of Differential Models with Applications to Parameter Estimation
- 2016-2018
- MSc, Mathematics; University of Calgary (Calgary, AB, Canada)
Hobbies
- Spoken Languages:
- I love jazz music, flight simulation, and cooking
contact@iliailmer.com Fredericksburg, VA