Algorithms, pipelines, systems.
MedDataMCP
Unified MCP server for federated medical/biosignal dataset discovery — one query across PhysioNet, NSRR, Zenodo, HuggingFace, OpenNeuro, and Kaggle, returning ranked, normalized results for LLM clients.
- Connector-based registry fanning out to 6 public sources; adding a new repository is one connector file plus one registry line.
- find_paired_datasets for modality-pair discovery (e.g. PSG + IMU) — built for sleep-staging and biosignal research workflows.
- Normalized DatasetRecord schema (id, source, title, modalities, license, access tier).
- In-memory TTL cache, per-source error isolation, and boundary-aware modality keyword inference.
- pytest suite for connectors, registry fan-out, and modality detection.
- github.com/Harshith292002/med-dataset-mcp
3D Point-Cloud Registration & Surgical Navigation
Implemented 3D registration, calibration, error-modeling, and distortion-correction modules; cut ICP runtime 40% with bounding-volume hierarchies and reduced EM-tracking positional error by 30%.
- Modular library with per-component unit tests and verification documentation.
- BVH-accelerated nearest-neighbor lookup for ICP, dropping per-iteration cost on 100k-point clouds.
- EM-tracker distortion correction modeled as a polynomial deformation field calibrated from a known phantom.
Pneumothorax Detection & Segmentation
Trained a multi-task U-Net with residual blocks on the SIIM-ACR pneumothorax dataset; achieved IoU 0.65 and 80% accuracy with a custom class-imbalance loss.
- Custom class-imbalance loss combining Dice + BCE with positive-class up-weighting.
- Evaluated on Dice and IoU; analyzed failure modes across subgroups to quantify model confidence.
- Geometric augmentations (elastic deform, random affine) tuned for chest X-ray morphology.
Deep Unrolled Networks for MRI Reconstruction
Implemented ISTA-Net and ADMM-Net for accelerated MRI reconstruction from 25% undersampled k-space; extended ADMM-Net with Squeeze-and-Excitation blocks for 1.14× speedup at 24 dB PSNR.
- Unrolled optimization networks with learned proximal operators for sparse MRI reconstruction.
- SE-block extension to ADMM-Net improved convergence speed without sacrificing reconstruction quality.
- Full pipeline: k-space data loading, training, PSNR/SSIM evaluation, and ablation studies.
COPD Diagnosis from Lung Sounds
Engineered an at-home diagnostic ML pipeline from smartphone-recorded lung sounds using feature extraction, filtering, and classification on the ICBHI dataset — 95% accuracy on clean data.
- Compared logistic regression, SVM, and random forest classifiers on extracted acoustic features.
- Planned deployment with noise cancellation and medical-diagnostic regulatory considerations.
- End-to-end pipeline from raw audio ingestion through feature engineering to evaluation.
Noise-Resilient Emotion Recognition
Improved noisy-audio classification accuracy from 18% to 51% using MetricGAN+ denoising and Wav2Vec2 features; gained an additional 20% via HuBERT and BERT multimodal fusion on the MELD dataset.
- Staged model integration: denoising front-end → robust speech encoder → multimodal fusion.
- MetricGAN+ for speech enhancement under additive noise before feature extraction.
- HuBERT + BERT fusion for joint audio-text emotion classification.
3D Slicer Inference Tooling
Shipped Qt-based extensions for 3D Slicer that orchestrate deep-learning inference, so researchers and clinicians can run models without leaving the imaging environment.
- Native Slicer module wrapping model loading, preprocessing, and inference behind a clean GUI.
- Used by lab researchers and clinical collaborators to run segmentation and synthesis models on patient volumes.
- Bridged research-grade ML into a workflow clinicians already use daily.
Respiratory Rate from ECG
A signal-processing pipeline that estimates respiratory rate directly from a single-lead ECG, achieving 2.5 bpm MAE on the BIDMC benchmark.
What I build.
I build end-to-end ML systems that turn clinical and physiological data into validated, shippable outputs — sleep staging and fragmentation from multi-modal biosignals, ECG respiratory-rate estimation, GPU-accelerated 3D imaging pipelines, and tooling that puts models in the hands of clinicians.
Currently an algorithms engineer at Neurava and a research assistant at Johns Hopkins ISTAR Lab. I care about rigorous evaluation harnesses, reproducible training configs, and decomposing fuzzy problems into components that actually deploy; with hands-on exposure to FDA-regulated medical device workflows.
What I reach for.
Where I've been working.
Wearable sleep staging via PSG-to-IMU distillation, IMU pose estimation, and team MLOps.
- Built a PSG-teacher to IMU-student knowledge distillation pipeline for wearable sleep staging on the DREAMT corpus; benchmarked classical biosignal baselines (Cole-Kripke, vanHees) against transformer-based sequence models with a rigorous evaluation harness across paired biological corpora.
- Developed IMU-based human pose estimation with clustering (K-means, GMM, HDBSCAN) on bicep-mounted sensor streams; authored the ground-truth labeling protocol for posture-class validation.
- Established MLOps practices for the team: experiment tracking, dataset versioning, reproducible training configs, and evaluation dashboards for fast iteration.
Cross-modality 3D synthesis, GPU vision pipelines, multimodal acquisition, and clinician-facing inference tooling.
- Built MAISI-US, a ControlNet-conditioned 3D latent diffusion pipeline adapting NVIDIA's MAISI foundation model to cross-modality ultrasound-to-CT synthesis. First reported adaptation of MAISI to a non-CT input modality; trained only the ControlNet branch on 9,068 paired US-CT volumes for low-cost domain transfer while keeping the VAE and diffusion backbone frozen.
- Designed a multi-metric evaluation protocol (PSNR, mutual information, gradient orientation agreement, LPIPS) against paired ground truth, with parameter sweeps on inference steps, ControlNet scale, and noise factor for reproducible characterization of model behavior.
- Built GPU-accelerated 3D vision pipelines (volumetric resampling, 3D-2D registration, DRR generation) on a CUDA-based imaging library; implemented pose regression with a custom Res-UNet for real-time 2D-in-3D visualization.
- Led multimodal data acquisition integrating Clarius and Alpinion ultrasound probes, Medtronic O-arm CBCT, NDI optical tracking, and scripted Universal Robot (UR3) sweeps to produce a paired calibrated 3D dataset for training and quantitative evaluation.
- Shipped Qt-based tooling for 3D Slicer to orchestrate deep-learning inference, putting research models directly into the hands of clinicians and researchers.
ECG-based respiratory estimation and accelerometer posture classifiers for a wearable monitoring product.
- Designed ECG-based respiratory rate estimation using CEEMDAN, discrete wavelet transforms, and peak detection, achieving 2.5 bpm MAE on the BIDMC dataset.
- Implemented and unit-tested accelerometer-based posture classifiers for a wearable monitoring product.
Course support for CNNs, segmentation, and multimodal registration.
- Designed and graded assignments on CNNs, segmentation networks, and multimodal registration.
- Mentored students on PyTorch implementations, training diagnostics, and evaluation.
ECG biosensor testing and optimization in hospital environments for FDA-approved wireless devices.
- Optimized ECG data collection and testing procedures in hospital environments for FDA-approved wireless biosensors.
- Contributed to A2 ECG biosensor design; gained hands-on exposure to medical-device SoC and embedded workflows.
GPA 3.8/4.0 · AI, Biomedical Signal Processing, Medical Imaging, Medical Device Design · Deep Learning, ML for Signal Processing, ML for Medical Applications, Computer Integrated Surgery
GPA 3.76/4.0 · Embedded Systems, Signal Processing, FPGA Design
Reach me anytime.
Open to AI/ML engineer roles in medical devices — biomedical signal processing, medical imaging, wearable AI, and production ML pipelines. Also happy to talk research collaborations.