Data Manager

ADHIRAJ YADAV

*** * ********** ****, *********, MD 21210

443-***-**** • *******.***@*****.*** • www.linkedin.com/in/adhirajyadav

SUMMARY

Graduated with a Master’s degree in Computer Science, specializing in Machine Learning from Johns Hopkins University. Experienced in a variety of data science and machine learning technologies, I am highly focused with an ability to meet deadlines and deliver results with flexibility. I have excellent research skills, fueled by a personal aptitude of solving problems and a hunger to learn. I am on an F-1 Visa and I am authorized to work in the US for any employer.

ACADEMIC BACKGROUND

Johns Hopkins University, Baltimore, MD January 2016 - December 2017

Master of Science in Engineering, Computer Science (GPA: 3.6)

Highlighted Coursework: Machine Learning, Big Data, Computational Genomics, Machine Translation, Algorithms

NMIMS University, Mumbai, India August 2011 - May 2015

Bachelor of Technology, Computer Engineering

Highlighted Coursework: Intelligent Systems, Data Warehousing and Mining, Operating Systems, Distributed Systems

TECHNICAL SKILLS

Programming Languages: Python, Java, C++, C#, SQL, Perl, OCaml, HTML/CSS/JavaScript, PostgreSQL

Machine Learning: Classification, Regression, Clustering, Random forest, Boosting, Decision Trees, Neural networks, SVMs, PCA, Deep Learning

Data Science/Big Data Tools and Frameworks: Apache Hadoop, Kubernetes, Tensorflow, PyTorch, SciKit, Pandas, Numpy, SciPy, OpenMP, MPI, Apache Spark, Git, Amazon Web Services, LaTeX, jQuery, Tableau

WORK EXPERIENCE

Johns Hopkins University: Research Assistant (Machine Learning Engineer) March 2017 – December 2017

Collaborated on the Project focused on Predicting Incidence of Cardiogenic Shock.

Parsed Johns Hopkins Hospital Medical data to populate the feature space for the purpose of model training.

Refined the existing ETL module to accommodate domain specific changes.

Configured the ETL module to run on a heterogeneous AWS cluster using kubernetes.

Generated quantitative and qualitative summaries for feature analysis using Python and PostgreSQL.

Designed a visualization tool in Python (graphviz, matplotlib, pandas) to analyze patient allocation between care units in hospitals.

Enhanced the existing Tensorflow based model to accept domain specific data.

Performed exploratory research in order to streamline patient throughput through early prediction of post-surgery rehab and subsequent intervention.

Johns Hopkins Pathology Department: Software Engineering Intern May 2016 – Oct 2016

Performed full stack updates to the web application cbioportal.org developed for the purpose of cancer research. The updates included the addition of clean and precise tabular views to better access and review the gene mutation data of the patients and streamlining the flow through the system.

Developed a standalone sample viewer displaying all relevant data for each sample including mutation events and protein changes for each event.

SIGNIFICANT PROJECTS

Intelligent Soccer Simulator (Machine Learning)

Implemented an algorithm using a multiclass SVM classifier enabling the manager of a soccer team to simulate an entire season for the team.

Compared results of various machine learning algorithms: Logistic Regression, Random Forest, Boosting, SVM, Neural Networks.

Supplied the manager with the ability to review squad decisions for an individual game as well as for the entire season based on the results provided by the simulator.

Comparing Gene Regulatory Networks by Cell Type (Machine Learning on Genomic Data)

Constructed gene networks corresponding to varying cell types using a graphical lasso/co-variance matrix approach

Compressed the feature space using PCA techniques and computed the Bayesian information criteria to optimize GMM clustering.

Viewing Ribosome Profiling Data from Protein Sequence Level (Genomic Data Analysis)

Developed an application to view Ribosomal profiling data from a protein sequence level as opposed to the traditional method of viewing it via DNA sequence level to provide a more direct view of the data.

Engineered a complex pipeline to process the raw data by removing the sequencing bias from the experiment and allowing for mutations so as to obtain the potential binding motifs of molecular chaperones.

Comparative Study: Parallel Graph Framework (Big Data)

Compared the performance of Apache Giraph which is a graph parallelization framework on varying density and graphical structures in a MapReduce paradigm using Hadoop.

Analyzed their benchmarks and highlighted their differences based on graph partitioning, memory organization and I/O overhead.

Streamster (Web Service)

Created a modular multi-tier black-box system that allowed for real-time transfer of data from publisher to subscribers through data streams.

Configured a RESTful server utilizing Java web services to manage the data transfer while handling subscriber payments that scaled on a heterogeneous cloud-based system.

Web Crawler: Comicrawler (Web Search Engine)

Developed a web crawler that retrieved relevant news interests from the world of comic books tailored to user preferences and using recommendation algorithms.

Introduced functionalities like free form text querying, character comparisons and popularity measure.

CERTIFICATIONS

Deep Learning Specialization by deeplearning.ai via Coursera.org Sept 2017

Learned the foundations of Deep Learning, building complex neural networks and leading successful machine learning projects.

Implemented deep learning on real world case studies with data from Healthcare, NLP, Computer Vision and Autonomous Driving.

Gained practical knowledge on the implementation and optimization of Conventional Deep Nets as well as variations of LSTMs, GRU’s and Convolutional Nets through application on datasets of differing domains.

Microsoft Technology Associate by Microsoft IT Academy Nov 2014

Contact this candidate