SIGN UP FOR FREE | SIGN IN
Thinkers360
Interested in getting your own thought leader profile? Get Started Today.

Prasanth Tirumalasetty

Rancho Mission Viejo, United States

I operate at the convergence of Deep Tech and Patient Safety, believing that the true value of AI lies not just in algorithms, but in its ability to ensure compliance and save lives.

Currently, I serve as a technical lead in the medical device industry, where I architect .NET Core-based systems that validate critical sterilization standards (ISO 17025, USP 85, ISO 11737). My work ensures that federal regulatory requirements are met through automated, data-driven integrity—transforming "compliance" from a checkbox into a proactive safety engine.

Key Contributions & Thought Leadership:

Innovation: Inventor of the "Forma Smart Recovery Mat" (Granted UK Patent No. 6488221), a novel device designed to accelerate patient recovery.


Research: Author of multiple peer-reviewed papers on cutting-edge AI, including "Generative Supply Chain Digital Twins" (Accepted by Springer, 2025) and "Automated Surgical Inventory Sterilization" (Accepted by IEEE, 2025). My earlier work on "DeepCpG" for genomic analysis was published in Membrane Technology .






Leadership: Elected as a Distinguished Fellow of the Soft Computing Research Society (SCRS)—a recognition bestowed upon the top 0.5% of the organization’s 14,000+ members.


Impact: Developed privacy-preserving synthetic data frameworks for Neovatic Technologies, reducing data processing error rates by 60% and ensuring GDPR/CCPA compliance.


Industry Voice: Regularly serve as a judge for global innovation competitions, including the Edison Awards and BIG Innovation Awards (90+ nominations judged) .

I am passionate about helping organizations navigate the complexity of AI adoption in regulated industries, ensuring that innovation never comes at the cost of safety or privacy.

Core Competencies (For Tags)
Healthcare Informatics

Generative AI & Synthetic Data

Regulatory Compliance (ISO/FDA)

Supply Chain Digital Twins

Computer Vision

Data Privacy & Machine Unlearning

Available For: Authoring, Consulting, Speaking
Travels From: Los Angeles
Speaking Topics: Agentic AI, Healthcare ,

Prasanth Tirumalasetty Points
Academic 25
Author 1
Influencer 0
Speaker 10
Entrepreneur 0
Total 36

Points based upon Thinkers360 patent-pending algorithm.

Thought Leader Profile

Portfolio Mix

Company Information

Company Type:
Minimum Project Size: Undisclosed
Average Hourly Rate: Undisclosed
Number of Employees: Undisclosed
Company Founded Date: Undisclosed

Areas of Expertise

Agentic AI 30.43
Agile
AI 30.01
AI Ethics
AI Governance
AI Infrastructure
AI Orchestration
Analytics
Architecture
Big Data 30.09
Business Strategy
Cloud
COVID19 31.34
Creativity
Data Center
Design
Emerging Technology 30.23
Energy 37
ERP 30.10
Generative AI 30.35
GRC
Health and Safety 32.61
Healthcare
Innovation
IoT
IT Leadership
IT Operations
Open Innovation
Quantum AI
RPA
Security 30.24
Supply Chain 30.15

Industry Experience

Automotive
Chemicals
Healthcare
Oil & Gas
Pharmaceuticals

Publications & Experience

1 Article/Blog
Implementation & Analysis of Coded Machine Unlearning Protocols
International Journal of Leading Research Publication (IJLRP)
April 20, 2022
There are some applications that may necessitate removing the trace of a sample from the system, such as
when a user requests that their data be deleted or when corrupted data is discovered. Simply removing a
sample from storage does not necessarily remove its entire trace, because downstream machine learning
models may store some information about the samples used to train them. If a sample is completely
unlearned, it can be completely unlearned. We retrain all models that used it from scratch, removing that
sample from their training dataset. When multiple such unlearning requests are anticipated, unlearning by
retraining becomes prohibitively costly. The training data can be divided into smaller disjoint shards and
assigned to non-communicating weak learners using ensemble learning. Each shard is used to create a
faulty model. These models are then combined to form the final central model. In this paper, we propose
a coded learning protocol in which the training data is encoded into shards prior to the learning phase
using linear encoders. In addition, we present the corresponding unlearning protocol and demonstrate that
it meets the perfect unlearning criterion.

See publication

Tags: AI, Big Data, Energy

2 Conference Publications
A GENERATIVE FRAMEWORK FOR SUPPLY CHAIN DIGITAL TWINS: SIMULATING AND OPTIMIZING LOGISTICS SCENARIOS USING AI-GENERATED DATA AND RULES
International Conference on Data-Processing and Networking (ICDPN-2025) (Springer)
November 08, 2025

See publication

Tags: Supply Chain

A Computer Vision and Machine Learning Framework for Automated Sterilization and Batch Validation in Regulated Surgical Inventories Warehousing
IEEE
October 23, 2025

See publication

Tags: Emerging Technology, ERP, Health and Safety

3 Journal Publications
A Data-Driven Modular Framework for Predicting Single-Cell DNA Methylation Landscapes
Membrane Technology
April 20, 2024

See publication

Tags: Big Data, Health and Safety

A Comparative Analysis of Regression Models for Predicting COVID-19 Mortality
International Journal of Intelligent Systems and applications in Engineering
December 25, 2023

See publication

Tags: COVID19

MSI-Multi-Step Interaction Networks for Spatial-Temporal Forecasting
International Research Publication and Journals
June 01, 2023

See publication

Tags: Security

1 Keynote
Data Synthetic: Using Generative AI to Augment Sales and Inventory Datasets for Enhanced Forecasting Models
Multidisciplinary International Journal
October 12, 2025
Retail forecasting often suffers from sparse observations, intermittent demand, promotion seasonality, and stockout
censoring—conditions that degrade the performance of both classical and deep forecasting models. We present a
practically oriented framework for data-synthetic augmentation: generating tabular and time-series records that
expand and rebalance training data for demand and inventory forecasts while preserving business constraints and
privacy. Concretely, we describe a pipeline that (i) models heterogeneous tabular covariates (prices, promos, holidays,
item/store attributes) with state-of-the-art generators such as CTGAN and diffusion models for tables; (ii) synthesizes
realistic multi-variate time series (sales, on-hand, shipments) using TimeGAN/DoppelGANger with conditioning to
respect calendars, promotions, and inventory non-negativity; (iii) trains forecasting targets with global models (e.g.,
TFT, DeepAR, gradient boosting, Prophet); and (iv) evaluates fidelity, utility, and privacy with a train-on-synthetic,
test-on-real (TSTR) protocol, membership-inference audits, and nearest-neighbor distance tests. We outline an
experimental design using the M5 retail benchmark and provide governance guidance (differential privacy, risk
scoring, and documentation) to operationalize synthetic augmentation safely. While we do not claim synthetic data is
inherently private, our framework shows how careful conditioning and formal privacy mechanisms can improve
model robustness, reduce cold-start errors, and de-bias rare events—without leaking sensitive records.

See publication

Tags: Agentic AI, Generative AI

Thinkers360 Credentials

5 Badges

Blog

Opportunities

Events

Add to Shortlist View Shortlist »

Contact Prasanth Tirumalasetty

Book Prasanth Tirumalasetty for Speaking

Book a Video Meeting

Media Kit

Share Profile

Contact Info

  Profile

Prasanth Tirumalasetty

Latest Activity

Search
How do I climb the Thinkers360 thought leadership leaderboards?
What enterprise services are offered by Thinkers360?
How can I run a B2B Influencer Marketing campaign on Thinkers360?