Padmaja Jonnalagedda Email & Phone Number

San Jose, Ca, Us

- Leveraging deep networks to synthesize novel protein conformations in molecular dynamic simulation process- Developed novel sampling protocol for generative modeling of Ab42 proteins under varying force fields- Extensive evaluation of realism in generated protein structures

- Developed SPACESeg and SPACESeg2.0, novel automated techniques for remote detection of potential signs of life on Earth via terrain cross-sectional images (funded by NASA Exobiology)- Developed a novel image representation technique (ARORA) capable of highlighting potential biosignatures from a limited dataset- End-to-end detection and characterization of desired structures - Classification and feature visualization using auxiliary image representation as a step towards translative research in detecting potential signs of life on other terrains such as Mars- Published a paper at IEEE TGRS (2024), CVPR AI4Space (2021, best presenter award), and LPSC

- Radiogenomic analysis of breast and brain cancers using histopathology and MR images- Developed novel generative models for end-to-end radiogenomic characterization of rare mutation (19/20 co-gain) in Glioblastoma patients, and the first deep-learning technique for modeling tumor invasion properties of Glioblastoma using limited data of 19/20 co-gain- Designed MVPNets for improved breast cancer classification performance while reducing resource utilization- Improved state-of-the-art tumor diagnosis by ~9% via novel data augmentation technique in breast cancer detection - Published papers in IEEE BIBE 2018, ICPR 2020, MIDL 2020, and IEEE TAI 2024

- Collaboration with the AFS team under the IARPA BRIAR project, leading model design for our team- In charge of data curation, evaluation, code contribution, high-priority deliverables, and team communication- Developed two novel body biometrics for complex, large-scale human recognition under extreme imaging conditions- Achieved the Phase 1 goal of the program within only 6 months, publications in IJCB 2023, and submitted a patent

San Diego, California, United States

- Worked on improving downstream OCR accuracy by detecting and correcting document distortions- Employed a latent spectral representation to model distortions with no prior assumptions and no paired ground truth - Proposed plug-and-play model does not have an upstream or downstream dependency with reversible transformations - Full control over distortion modeling statistics and choice of offline and online processing to preserve document resolution- Can correct multiple types and extent of distortion such as noise, blur, shadow etc. - Fast inference and noticeable improvement in OCR output using Tesseract engine - Patent # 11836972 accepted on December 2023

Cambridge, United Kingdom

- Part of InnerEye imaging team to study Glioblastoma (GBM) and Head-and-Neck Organs-at-Risk (OAR) - Designed optimal feature representation to integrate 3D context into 2D models for tumor extraction in GBM MRI scans- Designed pipeline to incorporate 2D context into 3D small crops using attention and localization for OAR assessment - Developed multi-task CNN framework for simultaneous domain transfer and segmentation between unpaired contrast and non-contrast liver CT scans for Cambridge Hackathon 2019 - Simultaneous domain transfer and segmentation between unpaired contrast and non-contrast liver CT scans using multi-task GANs

San Diego, California

- Assess the use of various sensors (IMU sensors like gyroscopes, accelerometers etc.) that could be mounted as body sensors on patients undergoing physical therapy- Worked in developing algorithms that could improve effectiveness in monitoring tele-rehabilitation systems- Performed design of experiments, worked with choice of sensors, data collection and choice of algorithms

- Masters thesis on "Augmenting subjective assessments with objective metrics in neuro-muscular disorders". Advisors: Dr. Harinath Garudadri, Dr. Tse Nga (Tina) NgCo-advisors: Dr. Andrew Skalsky, Dr. Leanne Chukoskie- Worked closely on development of instrumented glove to measure spasticity in patients with Cerebral Palsy, Traumatic Brain Injury etc.- Worked with doctors for data collection with real patients and design of experiments- Proposed a metric to assess spasticity using data from array of pressure sensors and inertial measurement unit, which was selected in IEEE HI-POCT 2016- Achieved high correlation (~87%) as an improvement on standard assessment metric- Received the Frontiers of Innovation Scholars Program (FISP) award grant in January 2017

Mumbai Area, India

- Suspect detection and recognition from CCTV video feed- Achieved face detection and recognition in low quality video feed even when faces cover 1% of the entire frame

Hyderabad Area, India

- Studied the missile seeker operations- Project on design and simulation of an algorithm for Improvement in Accuracy of Frequency Measurement in radar missile seekers- Proposed algorithms to improve measurement accuracy without altering hardware specifications based on wavelet decomposition and non-linear interpolation techniques- Achieved at least 25x more accuracy at noise level of about 0dB

Doctor Of Philosophy - Phd, Electrical And Computer Engineering

- GPA: 4/4 - Radiogenomic analysis of breast and brain cancers from microscopy and MR images - Automated protein synthesis for MD.

Master Of Science - Ms, Electrical Engineering

- GPA: 3.6/4 - Thesis on "Augmenting subjective assessments with objective metrics in neuromuscular disorders" - Graduate Student.

Bachelor Of Technology - Btech, Electronics And Communications Engineering

- GPA: 8.67/10 - Project on "Biometric devices to establish identity on campus: A Smart Campus Initiative" - Ladies Representative of IEEE.

Ssc