
ML Engineering at Kore Geosystems.
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Skills: Computer Vision, Machine Learning, Deep Learning, Image Processing, MLOps, LLMOps Experience: Object Recognition|| Detection|| Image Retrieval || Object Tracking || ReID || Image Classification || Key Point Detection || Semantic and Instance Segmentation || Text Recognition || LLM Engineering || Vision Language Models
Rochester Institute of Technology
Engineer's Degree, Computer Engineering
January 1, 2014 – January 1, 2016
University of Mumbai
Engineer's Degree, Electrical, Electronics and Communications Engineering
July 1, 2010 – January 1, 2014
KORE Geosystems
Lead ML Engineer
June 1, 2024 – Present
KORE Geosystems
Senior ML Engineer
June 1, 2022 – June 1, 2024
KORE Geosystems
Data Scientist
June 1, 2021 – June 1, 2022
G42
Computer Vision Data Scientist
September 1, 2019 – June 1, 2021
Abu Dhabi, United Arab Emirates
Blackstraw.ai
Data Scientist [Computer Vision/Deep Learning]
June 1, 2018 – June 1, 2019
Tampa/St. Petersburg, Florida Area
Entefy
AI / Machine Learning Engineer [Computer Vision/Deep Learning]
September 1, 2017 – February 1, 2018
San Francisco Bay Area
Kogentix Inc.
Jr. Data Scientist [Computer Vision/Deep Learning]
August 1, 2016 – August 1, 2017
Greater Chicago Area
Rochester Institute of Technology
Grader
August 1, 2015 – May 1, 2016
Rochester, New York Area
Rochester Institute of Technology
Computer Lab Assistant
January 1, 2015 – May 1, 2015
Classification of Hand Written Digits using Convolutional Neural Networks (Deep Learning)
February 1, 2016 – Present
> 5000 samples from the MNIST data set were used for the experiment. > 10 Layer network was used. > Cross-validation accuracy of 98.5% was obtained using 20 epochs.
Expression Intensity Recognition using Manifold Learning Techniques, Support Vector Machines and Sparse Representation
October 1, 2015 – Present
> Used the Cohn-Kanade data set of facial expressions for implementation of the model. > 6-basic expressions, i.e, Anger, Sad, Disgust, Happy, Fear, Surprise along with their intensities were classified. > Histogram of Local Binary Patterns (LBP) was used as features. > Dimensionality Reduction Techniques like PCA, Kernel- PCA and Locality Preserving Projections (LPP) and Linear Discriminant Analysis (LDA) were used. > Achieved accuracy above 90% using all the techniques mentioned above. > Locality Preserving Projections (LPP) gives the best result with an accuracy of 95 %. > Sparse Representation classification using Manifolds resulted in 93 % accuracy.
Facial Expression Recognition
September 1, 2015 – Present
-> Used the Cohn -Kanade data set of facial expression for implementation of the model. -> 6-basic expressions, i.e, Anger, Sad, Disgust, Happy, Fear, Surprise were classified. -> Histogram of Local Binary Patterns was used as features. -> PCA was used for dimensionality reduction and Support Vector Machines for classification. -> Achieved an accuracy > 90% for cross-validation over the data set.
Object Recognition using Histogram of Oriented Gradients as features and Support Vector Machines for Classification (Using Matlab)
September 1, 2015 – Present
> Histogram of Oriented Gradients (HOG) works very well as a feature for object recognition. > 5 classes( i.e Cars, Helicopters, Buses, Butterflies, Motorcycles) were used for classification from the Cal-tech 101 dataset > Average Cross-Validation Accuracy obtained was 85 % > Principal Component Analysis and SVM’s resulted in an accuracy of 90%
Face Recognition using (PCA / Modular PCA) as features and Support Vector Machines for Classification (Using Matlab)
May 1, 2015 – Present
In this project, eigenfaces which is a part of principal component analysis (PCA) are used as features. The advantage of using PCA is that it reduces the dimensionality of the features and requires less amount of processing time. For the classification, we have used Support Vector Machines along with Polylinear Kernel. We have compared the performances of SVM on the ORL database of faces and the Yale Data set. ->>An accuracy of 82.5% in case of the ORL database and an accuracy of 88.89% in case of the Yale database of faces was obtained. ->>We used a different technique called as modular PCA and the accuracy obtained was more than the one obtained by using PCA. An accuracy of 84.68 % in case of the ORL database and an accuracy of 93.89 % in case of the Yale database of faces was obtained by using Modular PCA. ->>For Classification purpose using the Support Vector Machines we have used the lib-svm library.
Image Retrieval using Histogram and Texture Features (Using MATLAB)
April 1, 2015 – Present
1.Used unsupervised techniques like Histogram matching and Euclidean Distances for image retrieval from an image dataset. 2.Used LAWS filters for obtaining feature vectors which can be used for Image retrieval. 3.Compared the performance of both the techniques.
Grayscale and Color Image segmentation using Unsupervised K-means Clustering (Using Matlab)
March 1, 2015 – Present
Segmentation of color image based on k-means clustering of its grayscale intensities. Segmentation of color image based on k-means color clustering. A generic algorithm was developed to segment images from k=2 to k=12 clusters.
Line Following Robot using PID algorithm
September 1, 2014 – Present
> Used PID algorithm and implemented it using a Neural Network. > Driven by an arduino UNO (Atmega 32) and an IR sensor array was used as a sensing element. > The PID was tuned to reduce the error level and decrease the overshooting across the track. > This project was the final project for the course Machine Intelligence. > Tools used- Arduino IDE.
Obstacle Avoider Robot
August 1, 2011 – Present
Accomplished a project on ‘Obstacle Avoider Robot’ in the year, 2011.This was the first embedded system based project which I worked on. I did it as a hobby project. Made use of an IR- Led and a TSOP receiver to implement the Robot. It intelligently senses any obstacle coming in its way and swiftly maneuvers to avoid the collision. It was made by making use of an Arduino Development Board which was embedded with an ATMEL’S ATMEGA 32
Running Kubernetes Cluster using Azure Kubernetes Service
Educative
June 24, 2026 – Present
Advanced RAG Techniques: Choosing the Right Approach
Educative
June 24, 2026 – Present
LLM Engineering: Master AI, Large Language Models & Agents
Udemy
June 24, 2026 – Present
Working with Containers: Docker & Docker Compose
Educative
June 24, 2026 – Present
Cultural Fit Analysis
The candidate's experience is heavily skewed towards Machine Learning, Deep Learning, and Computer Vision, which aligns well with roles requiring strong analytical and model development skills. However, the target role is 'Data Analyst', which typically focuses more on data extraction, transformation, visualization, and reporting, rather than advanced model development. While the candidate possesses strong analytical skills, the direct alignment with typical Data Analyst responsibilities (e.g., SQL, BI tools, A/B testing, dashboarding) is not explicitly demonstrated in the provided experience or projects. The certifications in Kubernetes, Docker, and LLM engineering show a proactive approach to learning new technologies, which is a positive cultural indicator.
Soft Skills & Operational Fit
The candidate's project descriptions are detailed, indicating a methodical approach to problem-solving and a focus on technical outcomes. The progression through various roles at KORE Geosystems suggests dedication and growth within an organization. However, without specific behavioral assessment data, it is difficult to fully assess soft skills like teamwork or stress handling.