Srikanth Tadisetty

408-***-**** United States ad3m57@r.postjobfree.com

PROFESSIONAL SUMMARY

Result-driven professional with passion for machine learning, computer vision, and data science. Expertise in pattern recognition, statistical modeling, and data analysis. Proven track record of solving complex problems and presenting insights for informed decision making effectively. Experienced in algorithm and pipeline development for solving complex, individualized problems with a strong attention to detail. EXPERIENCE

Researcher in Deep Learning Lab, Kent State University August 2019 - Present PhD degree focused on exploring practices improving 4 applications

• Spearhead the development and deployment of four advanced machine learning applications, significantly enhancing research capabilities in the Deep Learning Researcher Lab.

• Lead the design and implementation of deep learning environments, managing installation, configuration, and end-to-end deployment on FIJI/ImageJ and institutional servers.

• Collaboration with University Hospitals, Cleveland Clinic, NEOMED, and NIH, creating scalable, task-specific custom pipelines to meet diverse research needs in biology and mental health.

• Conduct research in Manifold Neural Networks, providing insights into the geometry of high-dimensional deep ReLU networks and optimizing network lengths.

• Demonstrate expertise in data manipulation and ML/DL solution development using PyTorch and TensorFlow, contributing to significant, long-term research projects.

• Hands on experience with a profound understanding of both foundational and state-of-the-art deep learning models, including CNNs, DNNs, RNNs, LLMs, and Transformers.

• Exhibit strong mathematical acumen in deep learning fundamentals, including calculus, advanced linear algebra, probability, statistics, and manifold theory. Part-time Instructor & Teaching Assistant, Kent State University August 2018 - Present

• Instruct six courses at both Master's and Undergraduate levels, encompassing subjects as Data Security & Privacy, Information Security, and Digital Forensics.

• Serve as a special topic lecturer in Deep Learning Statistical Foundations, teaching complex concepts such as Multivariable Calculus, Advanced Linear Algebra, Spectral Graph Theory, PCA and CNNs

• Play a key role as a grader for Discrete Structures, actively contributing to the academic development of students in foundational computer science courses. Researcher in AISP Lab, Kent State University August 2017 - August 2018 Master’s degree focused on practices improving 2 applications

• Led a project focusing on the prediction of psychosis onset from social media data using NLP and machine learning, achieving significant breakthroughs in mental health analytics.

• Collaboration with NEOMED to develop a custom corpus for prediction language criteria.

• Developed and managed an automated process for extracting, processing, and analyzing chaotic Twitter data.

• Gained comprehensive knowledge in statistical modeling techniques, including SVM, XGBoost, Random Forest, KNN, K-Means, and Naive Bayes, applying them effectively in research.

• Contributed to a project focused on limiting data access with a privacy protection layer application in Tizen OS.

• Exhibit knowledge is building lattice structures for data access and Android app deployment. Researcher in the Department of Orthopedics February 2014 - March 2015 Lady Davis Institute Jewish General Hospital, Montreal, QC

• Participated in groundbreaking research focusing on intervertebral disc degeneration and its implications for back pain, under the guidance of Dr. Fackson Mwale and P. Madiraju (MSc).

• Contributed to a study investigating the effects of Link N on nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), and substance (PSP) in human fibrosus cells and bovine discs.

• Assisted in isolating and culturing fibrosus cells from human lumbar spines, analyzing the impact of proinflammatory cytokines on these cells.

• Played a key role in the experimental process, which included the administration of Link N and assessment of its ability to reduce neurotrophin expression and NGF release in degenerating discs. CORE COMPENTENCIES

Programming Languages: Python Java C++ MATLAB JavaScript HTML CSS SQL Flask Pytorch Lightning Machine Learning Frameworks: PyTorch TensorFlow Keras Scikit-learn OpenCV CUDA NLTK LangChain Hadoop Spark

Data Analysis & Visualization: NumPy Pandas Matplotlib NetworkX Seaborn TSNE Tableau Excel Cloud: AWS - IAM, EC2, S3, Lambda, Glue, Athena and RDS Web Scraping APIs Travis-CI VectorDB PineCone FAISS Streamlit ReAct

Operating Systems: Linux Unix Windows macOS VM Project Management Tools: Git MSOffice Suite MS Project SharePoint PowerPoint JIRA Models: Deep Neural Networks Convolution Neural Network Recurrent Neural Networks Transformers Vision Transformers Large Language Models (LLMs) ImageJ Regression Classification EDUCATION

Kent State University, Kent, OH

PhD in Computer Science GPA 4.0

MSc in Computer Science GPA: 4.0

Concordia University Montreal, QC, CA - Graduate Diploma in Computer Science Concordia University Montreal, QC, CA – BSc Microbiology MACHINE LEARNING PROJECTS

Biological Sample Segmentation August 2022 - Present

• Implemented a specialized segmentation pipeline for distinguishing astrocytes from 3D projected Cuprizone model mouse CT scans.

• Developed a tool capable of accurately and efficiently segmenting astrocyte branches, cell bodies, and locating correct nuclei in multi-channel high-resolution microscopic images.

• Innovated a unique labeling strategy to facilitate training and maintain result consistency across diverse cell morphologies and overlapping objects.

• Achieved an 88% segmentation accuracy outperforming Segment Anything (SAM) and Fast Segment Anything (Fast-SAM) with metric visualization.

• Developed and implemented a specialized 3D pipeline to enhance the efficiency of accurate segmentation directly in 3D.

• Utilized graph embedding, K-core and optimal transport methods to enhance data connectivity across slices.

• Supplied biologists with data visualization and metrics providing insight tailored to their specific needs. LangChain LLM Interface March 2023 - May 2023

• Utilized LangChain as intermediary to interface with LLMs available on platforms such as Hugging Face and ChatGPT 3.5.

• Engineered an application capable of extracting information from Twitter and LinkedIn accounts based on provided name, including images, summaries, intriguing facts, and interests. Constructed using a combination of applied chains, custom agents, output parsers, and Flask framework.

• Implemented a query retriever incorporating vector databases (particularly Pinecone), embeddings, and RetrievalQA. Extended framework with the FAISS package for local vector storage. Subsequently, employed Streamlit to fashion a user-friendly chat interface, integrating memory functionality to recall past interactions in the chat or better subject querying.

• Engineered a slim version of a code interpreter using LangChain's functionalities. Optic Cup and Optic Disc Ratio in Glaucoma Patients August 2022 - March 2023

• Devised a specialized machine learning pipeline for segmenting the optic cup and optic disc in glaucoma patient derived data using instance segmentation.

• Achieved a breakthrough 96% segmentation accuracy, outperforming existing models on the same dataset with output data visualization.

• This level of precision significantly improves early glaucoma detection, demonstrating the impactful application of machine learning in medical imaging. Cloud-Based Real-Time Data Analytics Platform October 2021 – January 2022

• Worked on the creation of a Cloud-Based Real-Time Data Analytics Platform, enhancing skills in cloud computing and large-scale data handling.

• Integrated AWS services, including IAM, EC2, S3, Lambda, Glue, Athena and RDS, for efficient data processing, storage, and management.

• Utilized Python for scripting and SQL, Triggers, Spark and Glue Studio for database query automation focusing on real-time data processing and analysis.

• Conducted analysis using AWS QuickSight of Youtube data across various countries. Low Dimensional Geometry of ReLU Networks August 2021 - Present

• Illustrated hyperplane arrangement in both multi-dimensional and two-dimensional settings to elucidate neural network decision boundaries.

• Provided insights into the limitations of employing lengthy networks with straightforward data.

• Established a connection between the theory of oriented matroids and Deep Neural Networks (DNN) to enhance the understanding of the geometric process in ReLU networks. CNN Architectures August 2021 - January 2022

• Executed implementations of various CNN architectures using PyTorch, including LeNet, AlexNet, and ResNet, UNet, MaskRCNN, YOLO, etc.

• Explored advanced models like Vision-Transformers and BERT, contributing to the understanding of evolving CNN technologies.

• Enhanced skills in instance segmentation and language models. Classical Eigen-faces August 2019 – September 2019

• Implemented classical Principal Component Analysis (PCA) and Singular Value Decomposition (SVD) for facial recognition and image reconstruction.

• Developed a model to reconstruct facial images using linear combinations of eigenvectors from a low- dimensional subspace, showcasing the application of mathematical concepts in computer vision.

• This project highlighted the intersection of machine learning and image processing, advancing my understanding of facial recognition technologies.

Predictive Analytics for Consumer Behavior October 2017 – December 2017

• Developed a machine learning model to predict consumer purchasing trends based on historical sales data, achieving a predictive accuracy of 92%.

• Conducted comprehensive data cleaning and preparation, ensuring data integrity for effective modeling.

• Implemented feature engineering to identify key predictors in consumer behavior.

• Utilized a range of models including Random Forest, XGBoost, and Neural Networks, refining them to optimize predictive performance.

• Applied Python, Pandas, Matplotlib, and Seaborn for data processing, analysis, and visualization.

• The project's insights now inform targeted marketing strategies, demonstrating practical business applications of data science.

Anonymous Prediction of Mental Illness in Social Media August 2017 – August 2018

• Developed an Scikit-learn NLP-based model to predict psychosis onset in Twitter users, employing preprocessing, advanced data mining and lexical analysis.

• Unstructured data sourced from ~50000 posts with Twitter API and Beautiful Soup subjected to cleaning processes using regex and NLTK.

• Collaborated with a Psychiatrist to build a custom corpus, leading to an innovative process for flagging potential mental health issues across social media. Map Reduce Java Friends Suggestion February 2017- March 2017

• MapReduce program in Java to suggest potential friends within a social network, leveraging common connections.

• Utilized big data processing techniques to efficiently handle and analyze extensive social network data.

• Demonstrated expertise in distributed computing and data mining, enhancing user engagement and network connectivity through intelligent friend recommendations.

Contact this candidate