Dhiraj Sharma

Lead the AI team to build state-of-the-art predictive models for multiple disease conditions and formulate the use of LLMs for next-best actions. Provide strategic direction for medical AI initiatives.

• Lead the AI team at Delorean AI to build state-of-the-art predictive models to predict events and possible future treatments in multiple disease conditions, including, chronic kidney disease, dialysis, cardiac vascular disease, mental health, and COPD for dialysis centers, physician groups, insurers, skilled nursing facilities, and workers’ comp management companies.
• Formulate the use of LLMs

Continuation of the prior role with an extended team, including digital strategy and digital transformation.

• Lead AI and data science projects in multiple industries, including, financial services, banking, pharmaceuticals, manufacturing, consumer packaged goods, oil and gas, and HR.
• Model cancellations and delays of specialty pharma prescriptions; identify salient factors; identify patients group for further follow up and intervention.
• Built a document clustering system to match resumes that is qu

• Leverage data science, machine learning, natural language processing and big data technologies to help the firm make informed and objective people decisions.
• Resume screener for high-volume jobs such as software developer, financial planner and analyst. Created novel document matching techniques to sift out quality resumes (NLP, KNN analysis, similarity measures).
• Cluster past resumes so th

Lead Data Scientist – Paragon Development Systems (2016-present)
• Built knowledge data bases to store (a) body parts and symptoms and (b) diseases and symptoms. Designed high-level languages to express the two types of knowledge and wrote their translators.
• Built an inferencing engine to decide, using supplied symptoms and the knowledge databases, the most likely diseases and additional sympto

• Provide consulting services in the areas of retention and acquisition of customers and employees.
• Customer retention and acquisition models, customer-product affinity based on latent attributes of both, contact stream design to avoid fatigue, customer similarity models for co-marketing.
• HR technologies, employee retention analysis and modeling, factor analysis of surveys, machine learning ap

• Attrition modelling using logistic regression, CHAID trees, and similarity analysis.
• Factor analysis of employee survey data to determine employee engagement drivers.
• Text analysis of comments in survey responses to identify focus areas for improvement.
• Identify salient causes in free-form text from exit interviews.
• Extracting features from resumes for machine learning. Latent traits d

• Team lead for extracting features from resumes for machine learning. Developed a production system to rank resumes using TF/IDF, cosine similarity, and KNN techniques.
• Conceived, validated, and built systems to
• Identify specialized skills (such as technology, finance, etc.) based on dictionary-based ontology.
• Classify resumes to identify who to interview and for what divi

• Designed and developed parts of software to parse and store 20+M emails/chats/voice recordings per day with real-time compliance and integrity checking.
• Architected scalable search and analysis solutions on multiple nodes using synchronization and fault-tolerance similar to those in map-reduce systems.
• Created a system to audit the duplicated database of 2+B items that was spread across 20

• Built a non-SQL, columnar database engine in C for analytical applications, which is 50 times faster than conventional SQL systems, does not need any indexing, and provides 2:1 storage compression. It supports data loading, transformations, Q/C, reporting, analysis, and limited modeling in one system.
• Used this engine to provide data analysis services to clients through MegaPlex (my own small

• Served as technical lead on several analytical projects including campaign management, customer segmentation, name selection, code assignment, response analysis, and product affinity determination.
• Industries: retail, ecommerce, insurance, and financial services.
• Clients were premier direct marketing companies: Aegon Insurance, CUNA Mutual Insurance, Scudder Stevens and Clark, Discover Car

• Lead the architecture team for the GSM wireless base stations for Bell Labs/Lucent. Designed real-time software and several protocols to coordinate call setup, handoff, termination, etc. This system was developed by a team of 100+ developers. (1986-1988)
• Developed real-time code for channel assignment in wireless base station controller at Bell Labs/Lucent; developed kernel code in C and assem

• Efficient compression of video signals: Created algorithms for optimal compression of video signals under channel bandwidth constraints using dynamic programming, temporal modeling, and stochastic analysis.
• Early Unix work on 32-bit computers: Ported all Unix utilities to the new platform.
• Computer graphics: Created an algorithm for display of overlapping objects on bit-mapped displays.
• Co

Pursued doctoral studies in computer science at Caltech, focusing on advanced topics in the field.