I'm a B.Tech student at Manipal Institute of Technology with strong interests in Machine Learning, AI for Healthcare, and Open Source scientific computing. I enjoy working on research-driven projects that combine theory, systems, and real-world data.

• Machine Learning & Deep Learning
• AI for Healthcare and Biomedical Data
• Systems, Performance & GPU Computing
• Open-source Scientific Software

Languages: Python, C, C++, Java, SQL
ML/DL: PyTorch, TensorFlow, scikit-learn, NumPy, Pandas
Data & Systems: Awkward Array, cuDF, GPU-accelerated pipelines
Tools: Git, Linux, VS Code, Jupyter, LaTeX

Active open-source contributor with merged PRs focused on GPU compatibility and API correctness.

Highlights:

• Contributed to improving ak.to_cudf compatibility with cuDF ≥ 24.12 in response to column API changes
• Updated internal column construction logic for cuDF arrays
• Added and iterated on GPU-gated test coverage for cuDF-related code paths
• Improved axis parameter handling and documentation across high-level APIs

Contributions include code, tests, and documentation — aligned with real user-facing issues.

Currently working on a clinical similarity model using the MIMIC-IV dataset:

• Patient representation learning from structured EHR data
• Similarity computation for cohort discovery and downstream clinical tasks
• Emphasis on reproducibility, ethical data usage, and scalable pipelines

(Independent research project, non-Kaggle)

Research-oriented project exploring neural decoding pipelines:

• Large-scale neural signal processing
• Training and inference on NVIDIA A100 GPUs
• Focus on model efficiency, data preprocessing, and evaluation

• Knowledge Graph to Text Generation
  Co-author, IEEE MoSICom 2025
  Ontology-driven triple-to-text generation using LLaMA-based LLMs
  (Journal manuscript under review at IEEE OJCS)

• Panorama Video Generation (Research)
  AI-based panorama and video generation; documented in LaTeX
• Handwritten Digit Recognition (CNN)
  PyTorch CNN with dropout, pooling, and GPU-enabled inference
• Custom Shakespearean LLM
  Transformer model trained on The Merchant of Venice
• Smart India Hackathon – Triveni IoT Solution
  Digital farm management portal integrating AI dashboards and IoT analytics

• Google Summer of Code contributor
• Long-term research in AI + Systems
• Meaningful open-source contributions with real-world impact

• GitHub: https://github.com/X0708a
• LinkedIn: https://linkedin.com/in/aashirvad-bajpai-534137286
• Email: [email protected]