Skip to content
View aguptan's full-sized avatar

Block or report aguptan

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
aguptan/README.md

Amal Guptan

  • Biomedical engineer experienced in algorithm design and large-scale data analysis across various modalities, with a strong ability to abstract complex problems and select suitable computational modeling strategies.
  • Experienced in system-level design, pipeline planning, and integrating tools across multi-modal data streams.
  • Developed pipelines for motor control, image processing, statistical modeling, and signal analysis (physiological and fluid-based) using MATLAB and Python.

About Me

  • M.S. Biomedical Engineering, The George Washington University (2024)
  • B.S. Biomedical Engineering, University of Rhode Island (2022)
  • Current Role: Biomedical Data Engineer (Research Technician) at Children's National Medical Center (CNMC)

Technical Skills

Languages: Python, MATLAB, R
Frameworks & Libraries: PyTorch, NumPy, Pandas, SciPy, OpenCV, scikit-learn, Matplotlib, Bioconductor (R)
Signal & Image Processing: STFT, FIR/IIR Filtering, Artifact Removal, Feature Extraction (SIFT/ORB), Wavelet/Fourier Analysis
Machine Learning & Optimization: Vision Transformers, PCA, Clustering (DBSCAN, K-Means), Hyperparameter Tuning (Optuna), Cross-validation (K-Fold, LOO)
Workflow Tools: Git, Jupyter, VS Code, Anaconda, Spyder
Systems & Workflow Design: Modular pipelines, dataset scaling, automation, toolchain integration


Selected Projects

EEG-Based Delirium Classification

  • Designed and deployed a full EEG classification pipeline using MATLAB and Python, integrating STFT-based Feature Extraction and Vision Transformer-based deep learning.
  • Designed and delivered a scalable solution for multi-subject EEG analysis across large pediatric datasets in MATLAB. Currently implementing multiprocessing to speed up spectrogram-creation for about 150 pediatric patients.
  • Built a modular object-oriented Python framework to move and copy files using the shutil library, fine-tune and validate Vision Transformer models for pediatric delirium prediction, incorporating Youden’s J statistic, Leave-One-Out, and K-Fold cross-validation for patient-specific models.
  • Optimized the deep learning frameworks for efficient parallelization for fold-wise GPU training. Delirium EEG Repository

Tiling and Stitching Pipeline for High-Throughput Microscopy

  • Owned the design and implementation of a Python-based imaging automation system for the Olympus BX-63 microscope.
  • Integrated Pycro-Manager with Micro-Manager 2.0 and servo-controlled stage movement to replace commercial stitching tools, significantly reducing lab costs by $15K.
  • Building an SIFT/ORB stitching pipeline and regularized least-squares alignment; applied Optuna multi-objective HPO + K-fold CV with Pareto/inter-fold analysis to a dataset-aware auto-tuner for new datasets. 1. Hardware Repository 2. Stitching Algorithm Repository

Continuous Non-Invasive Blood Pressure Monitoring

  • Led algorithm design for a novel fluid-based system for continuous blood pressure monitoring, driving research, and implementation to define key physiological parameters.
  • Owned the design and deployment of signal processing and data-driven algorithms, including FIR/IIR & Notch Filtering, Oscillogram Generation, and Envelope-based Visualization
  • Helped with developing and debugging the code for efficient data acquisition (Arduino) by implementing validation statements and editing faulty code.
    BPP Repository

Teaching and Research

  • Teaching Assistant, Introductory Genetics (GWU, Fall 2023)
  • Assisted Dr. Julia Omotade with course instruction and mentoring.
  • Awarded a paid fellowship for academic excellence.

Certifications

  • Statistical Analysis of Genome-Scale Data Science – Cold Spring Harbor Laboratory
  • PyTorch, Deep Learning, and C++ Courses – Coursera (in progress)

How I use AI for Coding

  • Breaking down complex problems & system design

    • Requirements → modules:

      - Turn problem statements into a module maps (data loaders, preprocessing, model/train loops, evaluation, logging) for architecture exploration.
      
    • Architecture exploration:

      - Compare design options (e.g., LOOCV vs K-Fold, single vs multi-objective HPO, patient- vs electrode-level analysis) 
      
      - Draft pros/cons before implementation.
      
      - Produce stepwise build plans (prep → train → validate → analyze) to keep work scoped and reproducible.
      
  • Speeding up development

  • Refinement

    • Speed

    • Refactoring

    • Robustness

  • Aid with documentation


Contact

Popular repositories Loading

  1. MicroscopeScanTool MicroscopeScanTool Public

    GUI-based microscope scanning tool with live preview, white balance correction, and FIJI-compatible metadata export using Micro-Manager and Python.

    Python

  2. DeliriumEEG DeliriumEEG Public

    Deep learning pipeline for EEG spectrograms using Vision Transformers, with cross-validation, threshold optimization, and patient-level evaluation.

    Python

  3. Blood_Pressure_Project Blood_Pressure_Project Public

    MATLAB/Arduino toolkit for oscillometric cuff analysis, including filtering, oscillogram generation, and MAP estimation validated against typical blood pressure device

    MATLAB

  4. StitchingAlgorithm StitchingAlgorithm Public

    Dataset-aware image stitching framework using SIFT/ORB matching, affine optimization, and Optuna multi-objective tuning for high-throughput microscopy.

    Python

  5. aguptan aguptan Public

    Biomedical Engineer | Signal Processing, Machine Learning, and Computational Pipelines for Healthcare Data