System Status: Online

AI/ML Engineer | Forward Deployed Engineer
GxP, GenAI & Audit Intelligence

I am Vinay Kumar K.V, an AI/ML engineer with 10+ years experience delivering solutions at TCS, building NLP and GenAI agentic systems for regulated pharmaceutical environments.

vk_terminal_v2.1
SYSTEM LOG ON: 2026-07-18T00:00:00. PROFILE CORE V3.0 ACTIVE.
Type 'help' or click the quick action chips below to query profile database.
guest@vatsa_node:~$

Bridging Strategy and AI Systems

The AI Engineer Who Works Across Business, Risk, and Delivery

I build AI systems that operate in the real world, including audit-anomaly detection, hierarchical NLP extraction, and GenAI tooling in regulated pharmaceutical environments.

With 10+ years at Tata Consultancy Services, I have moved from infrastructure operations into product engineering, onsite leadership, and now full-time AI/ML delivery. I am comfortable translating complex regulatory and operational needs into production-ready models and experiments.

Core Professional Differentiator
What I bring that most AI engineers do not: the ability to work on both sides of the table, leading requirements conversations with QA, Regulatory, and Operations teams while also building the models, pipelines, and agent architectures behind the solution.

GxP Compliance

Expert in pharmaceutical validation standards. Experienced in authoring full URS, SRS, and UAT protocols.

Doctoral AI Research

Conducting DBA research in Walsh College on robust safety-critical autonomous systems.

Interactive Skills Deck
Click on any skill node to query system details:

Machine Learning & NLP

Hierarchical NLP ModelsAdvanced
Classification & ExperimentationExpert

Generative AI & LLMs

LLM-Based Agent SystemsProduction-Ready
Document Intelligence EvaluationExpert

Enterprise Systems & APIs

Python Automation & APIsSenior Developer
GxP Delivery EnablementAdvanced

Leadership & Stakeholder Alignment

Cross-Functional DeliverySME
Agile & Scrum LeadershipPSM I

VATSA AI System

Unified Five-Modality Robotics Architecture

Neural Scanning Engine

Our vision modality uses a modified convolutional feed to process live visual fields. Built for low-latency robotics contexts, it operates on raw image inputs to identify operators and manufacturing bins.

96.31%
CIFAR-10 Accuracy
22.0
Real-time FPS

Career Journey

A Decade of Technical Growth & Leadership

AI/ML Engineer - NLP & AI Agents

Tata Consultancy Services
Jun 2026 - Present

Product Engineer - Lead / Business Analysis & GenAI Prototyping

Tata Consultancy Services
2023 - Jun 2026

Product Engineer - Lead / Validation / Scrum / Business Analysis

Tata Consultancy Services
2019 - 2023

Systems Engineer - IT Infrastructure Backup Automation

Tata Consultancy Services
2015 - 2017

AI/ML Engineer - NLP & AI Agents

Tata Consultancy ServicesJun 2026 - Present
London, United Kingdom
NLP Location PredictionExperimentation PipelinesLLM AgentsGxP Delivery
Live Platform Delivery: Building ML and NLP components for an AI-agent system that flags process gaps and deviations carrying audit and penalty risk, surfacing anomaly signals for investigators to review.
Hierarchical NLP Location Model: Developing a hierarchical, NLP-based location-prediction model to extract structured location signals from unstructured records.
Experimentation Infrastructure: Built a YAML-driven experimentation pipeline so features, datasets, and model choices can be swapped without touching code.
LLM-Based Agent Extension: Extending the system with Python-based LLM agent architecture for richer anomaly triage and investigation workflows.

LearnWithVinay

Founder & Tech Educator

LearnWithVinay is an ongoing personal education initiative. Currently, only Python-based material is in development; no C modules have been released yet.

Generative AI & LLM Systems

In Development

Advanced structures for semantic extraction, vector search indexes, vector db configurations, prompt safety filtering, and model benchmarking.

LangChainAdvanced RAGEvaluation MatrixAgentic Workflows
Syllabus progress

Machine Learning & Deep Learning

In Development

Mathematical backpropagation, gradient descents, custom neural classifier loops, and validation checks using SHAP and LIME.

PyTorch CoreCNNs / RNNsModel ExplainabilityValidation Metrics
Syllabus progress

Mathematics for Data Science

In Development

The fundamental mathematical building blocks behind regressions, loss functions, optimizer derivations, and neural weights.

Linear AlgebraCalculusProbabilityVector Projections
Syllabus progress

Python Programming

Recording In Progress

Writing clean, modular Python microservices and utility scripts designed to integrate into automated pipelines and live production environments.

OOP PrinciplesProduction ScriptsFile UtilitiesData Structs
Syllabus progress

SQL & Relational Databases

In Development

Relational database mechanics, constructing high-performance analytical queries, indexing methods, and transactional safety.

Indexed QueriesDB OptimizationAnalytical JoinsDB Schema Design
Syllabus progress

ROS2 & Robotics Systems

In Development

Robotics OS mechanics, wiring publisher-subscriber nodes, handling raw spatial lidar data feeds, and mapping joint motion instructions.

Nodes & Pub/SubSensory BindingsActuator ControlsSLAM Nav
Syllabus progress

Embedded C++ & PCB Design

In Development

Writing low-latency C++ instructions for microcontrollers, designing schematics, routing tracks, and ensuring clean hardware signal loops.

Firmware DevSignal IntegritySchematics RoutingMicrocontrollers
Syllabus progress

Academic Foundation

Combining Doctoral Research, Applied AI, and Business Rigor

Featured Publication

VATSA: Video, Audio, Text, Sensory, Action - A Unified Five-Modality Architecture for Human-Level Perception and Action

Authors: Vinay Kumar .K .V

Abstract: This research proposes a unified multimodal neural network model tailored for safety-critical robotics environments. By combining visual bounding boxes, audio spectrogram converters, text goal parsers, and spatial sensor networks, the architecture drives real-time action instructions. The model integrates the SAMOS routing mechanism, analyzing collision scores dynamically to restrict motor torque velocities in proximity to human operators, providing a robust validation pathway for GxP-regulated deployment.

Concept Paper

Doctor of Business Administration (DBA) in AI & Machine Learning

Walsh College, USA (via Great Learning)

2025 - Dec 2028Remote / US-based

Focus: AI & Machine Learning, with research interest in autonomous systems in safety-critical environments.

Post Graduate Program in AI & ML

McCombs School of Business, UT Austin

In progressRemote / US-based

Focus: Applied AI & ML with a business-oriented lens for real-world deployment.

Master of Business Administration (MBA)

MSRIT, Visvesvaraya Technological University

2017 - 2019Bengaluru, India

Focus: Marketing, with a strong grounding in business strategy and stakeholder communication.

Bachelor of Engineering (B.E.)

Sir M. Visvesvaraya Institute of Technology

2012 - 2015Bengaluru, India

Major: Electronics & Communication Engineering.

LiveConnected to the Hugging Face digital twin.

Ask me anything about Vinay's digital twin, portfolio, or recent projects.