Harsha Pokkalla
Data AnalystVP of AI/ML @ PathAI | IITG, CMU
PathAI
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
VP of AI/ML @ PathAI | IITG, CMU
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Building domain-specific multimodal foundation models and AI products for histopathology. Our products support biopharma translational research, clinical trials, and digital diagnostics customers globally.
Carnegie Mellon University
Master's in Electrical and Computer Engineering, Machine Learning and Computer Vision
January 1, 2015 – January 1, 2016
Indian Institute of Technology, Guwahati
Bachelor of Technology (B.Tech.), Electronics and Communications Engineering
January 1, 2011 – January 1, 2015
PathAI
VP of Machine Learning R&D
November 1, 2023 – Present
PathAI
Director of Machine Learning
May 1, 2021 – October 1, 2023
PathAI
Employee #4 / Sr Machine Learning Scientist
February 1, 2017 – April 1, 2021
NEC Laboratories America, Inc.
Summer Research Assistant
May 1, 2016 – August 1, 2016
Princeton, New Jersey
Carnegie Mellon University
Graduate Research Assistant
January 1, 2016 – December 1, 2016
Nav Lab, Robotics Institute
ETH Zurich
Summer Research Fellow
May 1, 2014 – July 1, 2014
Zurich, Switzerland
LVPEI (collaboration with Camera Culture group, MIT Media Lab)
'Engineering The Eye' Workshop, MIT Media Lab
July 1, 2013 – July 1, 2013
Greater Hyderabad Area
Indian Institute of Technology, Hyderabad
Summer Research Fellow
May 1, 2013 – June 1, 2013
Greater Hyderabad Area
Relationship Verification using Siamese Networks
April 1, 2016 – Present
- Using siamese networks, built deep CNN model to verify kinship relationships (father-son, father-daughter, mother-son, mother-daughter) using face images. KinFaceW dataset was used for experiments. Achieved 79.3% overall accuracy. - Extended the same idea for couples by downloading celebrity couples in the wild. Similar looking couple faces are clustered together in deep feature representation space.
Yelp Restaurant Photo Classification using CNN features
March 1, 2016 – April 1, 2016
Using pre-trained Alexnet model, extracted the fc7 features of the given images and build multi-label SVM Model for classifying restaurant images.
Implementation of Deep Learning Techniques in Caffe & Torch
March 1, 2016 – April 1, 2016
- Object classification and object localization on MS COCO Dataset (fine tuning the Alexnet and training network from scratch). - Joint learning for object classification and object localization through single network on MS COCO Dataset. - Nearest neighbour visualization of CNN pool5 and fc7 features for object classification and localization models. - Surface normal estimation on NYUv2 dataset using Fully Convolutional Networks in Torch (fine-tuning Alexnet and training from scratch)
Localization of objects using pre-trained Convolutional Neural Networks
October 1, 2015 – March 1, 2016
- Using Pre-Trained CNNs, observed the variations in the output probability of the true class by masking the regions of the input image and activation maps at all convolutional layers. - Proposed techniques to find the regions of input image that generate maximum responses towards the classification. Bounding box is generated over the object in the input image using these localized regions. - Implemented these techniques in Torch by loading AlexNet Caffe model with ImageNet Dataset as the testing dataset.
Face Recognition in Broadcast News Videos
September 1, 2014 – March 1, 2015
- Employed Face TLD tracker using Viola Jones method for extracting the face-tracks. - Generated feature vector by carrying out spherical clustering on collected face-tracks at multiple patch levels hierarchically. - Implemented One versus All SVM Classification technique for recognizing the faces on testing dataset.
Humanoid Face Bot - Mimics the facial gestures in real time
March 1, 2014 – Present
Designed a humanoid face bot that mimics real time face gestures (eye blink and jaw movement) using OpenCV and Arduino.
Pattern Recognition and Machine Learning Techniques
February 1, 2014 – April 1, 2014
- Implemented Principal Component Analysis for face recognition and reconstruction (Eigen Faces). - Otsu’s Image Segmentation for foreground and background separation. - Gaussian Mixture Model in image clustering scheme and object classification.
MedAssist - Automated Medication Kit for elderly people in India
November 1, 2013 – January 1, 2014
Developed an automated device that stores medicines, tracks prescription schedules, reminds and records patient’s responses.
Amsler Grid Test android app
July 1, 2013 – Present
- Developed an android application which will help as diagnostic tool for Age-related Macular Degeneration (AMD). - Takes convenient 'Amsler grid eye test' that can be carried on at any place to monitor the changes in the vision, especially for older adults.
Circuits and Electronics
edX
June 24, 2026 – Present
Digital Signal Processing
Coursera
June 24, 2026 – Present
Cultural Fit Analysis
The candidate's extensive experience in AI/ML, particularly in the medical domain, aligns well with roles requiring innovation, rigorous validation, and real-world impact. Their background in leading research and development, coupled with a history of building and deploying products, suggests a results-oriented and proactive approach. The diversity of personal projects, though older, indicates a strong foundational interest in various ML/CV applications. The candidate's progression within a single company (PathAI) from Sr ML Scientist to VP suggests loyalty, growth mindset, and ability to thrive in a challenging environment. While the target role is 'Data Analyst', the candidate's profile is significantly more advanced, indicating a potential mismatch in scope, but a strong fit for a senior ML/AI leadership role.
Soft Skills & Operational Fit
The candidate's career progression from individual contributor to VP of Machine Learning R&D at PathAI demonstrates strong leadership, strategic thinking, and project management skills. Their experience in navigating complex regulatory requirements (FDA/EMA) indicates a high degree of operational acumen and attention to detail. The descriptions of leading teams and managing roadmaps suggest excellent communication and collaboration abilities, crucial for a senior role. The candidate's early involvement as Employee #4 at PathAI also points to an entrepreneurial spirit and adaptability.