cv
An extended version of my resume. For a shorter one, please refer to the pdf on the right.
Basics
Name | Adithya Narayan |
Label | Machine Learning Engineer |
anaraya2@andrew.cmu.edu | |
Phone | (412) 737-4602 |
Url | https://adithyaknarayan.github.io/ |
https://www.linkedin.com/in/adithya-n-b637b8146/ |
Work
- 2024.09 - Present
Research Collaborator
CMU Lab4D++ (Opensource)
Exploring generative image priors and sim-2-real pipelines for 4D reconstruction and interaction modeling.
- Investigated the use of generative image priors for Gaussian Splatting in 4D reconstruction tasks.
- Built a Unity-based sim-2-real pipeline for modeling human/scene interactions.
- 2024.09 - Present
Graduate Research Assistant
Human Sensing Laboratory | Meta Reality Labs
Collaborating with Meta to enhance the robustness of SoTA depth estimation and semantic segmentation models.
- Developed adversarial sampling techniques for 3D scene parameters using differentiable renderers.
- Created a synthetic data pipeline in Blender to animate trajectories and generate texture and depth maps.
- 2023.03 - 2024.07
Machine Learning Engineer
Arintra
Improved medical LLMs and system efficiency for clinical applications.
- Applied Retrieval-Augmented Generation to medical LLMs, improving ICD code prediction accuracy by 6%.
- Designed a CI/CD system for ML models using MLFlow, FastAPI, and GCP, ensuring zero downtime.
- 2022.02 - 2023.03
Machine Learning Engineer
Klothed
Worked on 3D reconstruction and synthetic data pipelines for consumer avatar generation.
- Integrated sparse-view 3D reconstruction models with diffusion inpainting for clothed avatar generation.
- Optimized 2D image warping pipelines using FEM techniques and SciPy, reducing inference time by 90%.
- 2020.11 - 2021.11
Research Engineer
Origin Health
Developed domain-specific synthetic data and segmentation pipelines for fetal brain imaging.
- Co-authored research improving fetal biometry measurements with synthetic data and attention mechanisms.
- Generated synthetic data samples using GANs and domain adaptation, boosting segmentation mIOU by 20%.
Education
-
2024.09 - 2025.12 Pittsburgh, PA
-
2017.07 - 2021.07 Karnataka, India
Publications
-
Validation of a deep learning system for the automated segmentation of fetal brain structures
ISUOG 2021
Validated deep learning system for fetal brain segmentation across devices.
-
Device-independent deep learning system for automated segmentation of fetal brain structures
SPIE 2022
Achieved device-independent segmentation for fetal brain structures.
-
Biometric Constraint Attention Masks to obtain Sonographic Measurements
IEEE ISBI 2022
Developed attention mechanisms for improving fetal biometry measurements.
Skills
Programming Languages | |
Python | |
C++ | |
Java | |
Bash |
Libraries and Frameworks | |
PyTorch | |
PyTorch3D | |
Torch-TensorRT | |
OpenCV | |
TensorFlow | |
Keras | |
Pandas | |
SciPy | |
NumPy |
Tools and Platforms | |
Docker | |
GCP | |
Redis |