- Data scientist from big data analytics perspective, focusing mainly on healthcare/pharmaceutical discoveries (cancer drug discovery).
- Scientist, with expertise in computer science, computational biology, biochemistry, toxicology studies, dietary antioxidants/human cancer risk assessment (genetic/environmental risk factors - diet, lifestyle patterns, behaviors).
- Ability to solve complex problems (diagnostic errors, medical errors/ADR) using AI, machine learning/data mining-with the goals to improve clinical outcomes in clinical practice/clinical research (preclinical/clinical studies, drug safety). Interest in precision medicine, pharmacogenomics & cancer care.
- Skill in leveraging integrated clinical big data (proteomic/radiogenomic/genomic/clinical data); data preprocessing, predictive models for disease target across varying patient cohorts using cleaned large dataset; feature extraction/selection-classification (build patient classification models) or train a linear regressor. On the basis of different AI methods (DNN, SVM, RF) for predicting patient clinical outcomes (diagnosis, prognosis)/survival.
- Skill in text data analysis through pattern analysis, text mining and NLP, derive insights and creation of cancer “knowledgeable/predictive engine” that can inform clinicians about best possible patient treatment options.
- Skill in data science/computational prediction method to solve complex public health problem (cancer) using big data, unsupervised (clustering algorithm)/supervised learning, deep learning (CNN, LSTM)-prediction tasks.
- Skill in toxicology of acrylamide (neurotoxicity, toxicokinetic, carcinogenicity/genotoxicity). Including chemical/drug metabolism; pharmacokinetics studies for clinical research (R &), dosing and precision medicine.
- Familiarity of public databases: PK-DB, OFFSIDES, Drugs@FDA/SIDER/SEER/DrugBank/PharmGKB-train text mining for pharmacogenomic studies, drug-drug interactions, drug-gene relationship, ADRs.
- Skill in computer science, software engineering (AI, NNs), computational biology-probabilistic model (Bayesian network), simulations, deep learning for NLP-EHR (clinical note-text, medications, ICD), FHIR, HIPAA.
- Familiarity big data technologies (NoSQL, cloud), data visualization/interpretation/presentation/communication (oral/written). Ability to collaborate with clinicians, make decisions, improve patient outcomes and safety.
- Skill in deep learning with Keras/TensorFlow & analytical tools: Weka, R/Python/Spark for big data processing, automated real-time analytic (cancer care, mobile health sensors/patient monitoring, ICU/home care).
- Skill in statistical inference, bioinformatics tool (Rainbow, GAKT)/molecular diagnostics (RT-PCR, NGS)-DNA seq/variant calling/SNP, pharmacogenomic test (cancer, mental health/pain medication insight), biomarkers.
- Skill in genetics study (genes or protein targets)/role of genetic factors in preventing disease, understands how association between genomic variation/drug response are investigated; predict the likelihood of side effects.
Confidential, Bethesda, MD
Healthcare Data Analyst
- Predictive analytics, healthcare fraud detection and prevention methods, using R/SAS Enterprise miner/machine learning models to detect abnormal patterns in Medicare data sets/conduct advanced data analysis-extraction of providers information from CMS-IDR warehouse for advanced analytics-upcoding (ICD, CPT billing code errors, univariate/multivariate outliers/anomaly detection model, anomaly scores. Bayesian network and data predictive models, flag potential fraudulent providers, train neural network models for anomaly detection-payment amount deviation and data points related to procedures performed/drugs administered.
- Use of autoencoders, unsupervised Random Forest/K-Nearest Neighbors, Local Outlier Factor, prevent healthcare fraud /waste/abuse. Medicaid/Medicare processing/frequency of reimbursement-detection/prevention.
Confidential, Rockville, MD
Sr. Pharmacy Data Analyst
- Experience in healthcare/pharmacy datasets (Rx claims/clinical data)-extracting prescription drug (Rx) information from Netezza data warehouse for analytics (ICD, CPT billing codes and drug utilization review tasks).
- Statistical data analysis approaches (SAS enterprise miner/Guide), outliers’ detections. Exploratory data analysis/clean data, data summaries via descriptive statistics and use of predictive analytics, drug mix, dosage, Rx counts/frequency claim/patterns. Log reimbursement Rx claims frequencies that looks suspicious. Use of NLP billing codes/medical claims (CPT/ICD-9/10/specialty drugs (e.g., for cancer), regression analysis/modeling.
- Applied machine learning algorithms (SVM, NN), train binary classifiers/estimate individual risk of high-cost claim/risk analysis-cost effectiveness models. Use R/SAS/Tableau for data visualization/reporting. Use in house cost models, retrospective drug (brand/generic drugs) utilization patterns, frequencies of ADRS/high cost patients review-therapeutic abuse by beneficiaries on pain medications in chronic care (cancer, mental health).
- Concern is that acrylamide, an environmental chemical (food contaminant that causes cancer in lab mice) may cause cancer in human.
- Cancer is a complex public health problem.
- Investigated acrylamide level in food and potential exposure to humans.
- Conducted experimental design, statistical data analysis and modelling of dietary nutrients (dose) required for the studies.
- Conducted adult population food consumption survey (dietary intake/lifestyle of adult individuals), measured and interpreted associations, collected data was combined with acrylamide toxic levels discovered in processed food items (toxicology lab, mass spectrometry, acrylamide levels in human diet, its metabolism and toxicity were measured), generated datasets were cleaned/summarized for risk assessment studies.
- Applied probabilistic model, variability in acrylamide exposure, estimated the likelihood of adult exposure to acrylamide (a chemical carcinogen), reasonably anticipated to be a human carcinogen. Known carcinogen in laboratory animals (mice) based on different published scientific research studies.
- Used R statistical tool in this research project for data analysis and visualization (charts, graphs), accurately interpreted data, reviewed outcomes/presented results to a group of scientific audience.
- There are no associations between dietary intakes (acrylamide) and risk of developing cancer by adult individuals.
- Taught graduate level courses: information security and database management systems
Confidential, Spring field, VA
- E-commerce transaction monitoring in data center environment, weblog data analysis and documentation, access control and validations.
- Investigated data security issues, use instruction detection systems, vulnerabilities and risk analysis, flag unusual transaction activities.