- 🎓 Master’s student in Artificial Intelligence @ Iowa State University
- 🤖 Passionate about Machine Learning, NLP, Deep Learning, and real-world AI systems
- 🛠 Recently been trying to build things from scratch — tools, pipelines, small products, experiments
- 📈 Interested in automation, XAI, efficient ML systems, and data-driven decision making
- 🥊 Outside tech: boxing, soccer, anime, music, and trying new cuisines
- Real-time nanopore sequencing analysis and visualization pipeline
- Architected snapshot-based incremental processing to enable efficient real-time analysis of streaming nanopore data without reprocessing previously analyzed files
- Created real-time interactive Plotly dashboard with multi-barcode visualizations (quality distributions, coverage plots, mapping statistics) and real-time updates
- Benchmarks and compares different NLP approaches for text analysis. Project inspired by my first NLP couse @ ISU (5790).
- Comparing rule-based techniques (regex tokenization), transformer models (BERT), and LLMs for sentiment analysis, topic classification, and summarization.
- Check the trade-offs between speed, accuracy, and complexity across these different methods with visualizations.
- Auto-renames, organizes, and tags documents
- PaddleOCR + Llama 3.1B pipeline, built for 50–100+ client firms
- Hardware + SaaS hybrid MVP
- Analyzing neuron activations in code-trained LLMs
- Clustering learned concepts
- Working on visualization + interpretability tools
- Efficient training & optimization
- Better engineering practices for AI projects
- How to design clean ML pipelines
✈️ From Nepal → Iowa- 🧐 Curious about how things work under the hood
- 🧪 Enjoy experimenting, improving, and building useful tools
- 🎯 Goal: become really good at ML
- 💼 LinkedIn: linkedin.com/in/manjulbalayar
- ✉️ Emails:
- Personal: [email protected]
- School: [email protected]