PhD researcher building ML systems that ship.
Nader Lobandi
Ph.D. Candidate · M.Sc. Data Science
Data Scientist & PhD Researcher
University of Denver
Machine Learning Engineer and PhD Researcher with 4+ years of experience designing, deploying, and scaling production-grade AI/ML systems across industry and academic research. Expertise spans agentic LLM applications, deep learning, real-time computer vision, and large-scale data pipelines on Azure and Databricks. Holds an M.S. in Data Science from Northeastern University with strong foundations in statistical learning, probabilistic modeling, optimization, and algorithm design. Seeking ML Engineer roles focused on building end-to-end intelligent systems and advancing applied AI and foundation model technologies.
Experiences
Data Scientist
BeanSprout AI
- ▸Built a cardiovascular risk prediction model using LightGBM and XGBoost, achieving AUC 0.89, with full lifecycle support from data ingestion through hyperparameter tuning to Modal deployment with MLflow tracking.
- ▸Implemented real-time voice interaction via Vertex AI STT/TTS and designed 3 agentic workflow types — secure document access, intent routing, and adaptive user support — for an elder-care platform.
- ▸Engineered a RAG pipeline over patient documents with persistent chatbot memory and user accounts, reducing repeated data-lookup steps to ~0 per session (~estimated).
Data Scientist
VHB (Vanasse Hangen Brustlin Inc.)
- ▸Analyzed terabytes of city-scale traffic data using PySpark, SQL, and Delta Live Tables on Databricks, uncovering that over 70% of streaming feeds originated from a single unreliable vendor.
- ▸Deployed an internal RAG chatbot with SQL execution, giving project managers natural-language access to employee hours, assignments, and reporting data in ~seconds vs. multi-day manual lookups (~estimated).
- ▸Developed LSTM, Prophet, and SARIMA time-series models to forecast energy demand for NY Transit Authority, improving load-prediction accuracy to support peak-reduction strategies.
Data Scientist Co-op
Kibur Medical Inc.
- ▸Built a bilateral filtering preprocessing pipeline for high-dimensional biomedical images, improving cell quantification accuracy by ~20%.
- ▸Integrated AWS S3 for scalable, real-time image retrieval, replacing local storage and supporting a ~3-month continuous clinical imaging data workflow.
Research Scientist
Northeastern University
- ▸Trained LSTM, Gaussian mixture, and regression models on time-series COVID biosensor data, achieving a best detection recall of 95% under noisy measurement conditions.
- ▸Published 4 peer-reviewed papers in IEEE Sensors Letters, Biosensors and Bioelectronics, and IEEE CICC on ML-driven biosensor detection systems.
- ▸Formulated hypothesis-testing frameworks using MAP, Bayesian, Neyman-Pearson, and Minimax criteria, reducing false-alarm trade-offs across 3+ detection threshold configurations.
Education
Ph.D. in Computer Science
University of Denver, Ritchie School of Engineering & Computer Science
FocusDeep Learning, Computer Vision, Foundation Models, GPU Computing & CUDA Programming
M.Sc. in Data Science
Northeastern University, Khoury College of Computer Sciences
HonorsGordon Institute LEADERS Fellowship; LEADERS Leadership Certificate
B.Sc. in Electrical Engineering
Sharif University of Technology
HonorsNational Physics Olympiad Silver Medal; Ranked top 0.1% out of 100,000+ applicants in national university entrance exam
Projects
Financial Time-Series & Document QA
Built an educational finance tool to help non-expert users explore forecasting strategies and consume company fundamentals from SEC filings in plain language. Enabled interactive backtesting with reliable forecast horizons up to two months, paired with a chat-based interface for querying 10-K reports.
Personal Finance Web App
Built a self-hosted alternative to tools like Intuit, giving users full ownership over their financial data while tracking budgets, expenses, and investment goals. Delivered dashboards and reports summarizing financial health, with automated alerts for suspicious transactions and secure multi-user authentication.
Financial News Sentiment Analysis
Built an NLP system to interpret high-volume financial news and quantify market sentiment, supporting faster and less biased trading-signal extraction. Benchmarked classical and transformer-based approaches across two datasets (Alpha Vantage news, Yahoo Finance), with FinBERT achieving 0.73 accuracy / 0.74 macro-F1 on Alpha Vantage and BERT-uncased reaching 0.86 / 0.86 on Yahoo — validated against real-world ticker movements (e.g., VLN, HSII).
E-Waste Sustainability Assistant
Built a vision tool that identifies electronic waste and returns its carbon footprint, recoverable value, and a safe disposal action, addressing the growing volume of recyclable e-waste in data centers. Delivered an end-to-end pipeline that turns a photo into a defensible refurbish/recycle/review decision in seconds, backed by a curated knowledge base with full source provenance for every CO₂ and metals figure.
Skills
By Domain
Machine Learning
Data Engineering
AI & LLMs
Proficient
Building Toward
Contact
ML Engineer and PhD Researcher at the University of Denver — building production AI systems at the intersection of deep learning, computer vision, and large-scale data. Open to collaboration and opportunities in ML Engineering, AI Engineering, and Data Science. If you're a builder or thinker working on hard problems, let's connect.
Ph.D. Candidate · M.Sc. Data Science · University of Denver