Shiv Prakash
Engineering Leader & Architect
Distributed Systems · Data · AI/ML
Graduate Researcher
Georgia Tech · ML · DL · RL · NLP · CV · Robotics · NSc · Quantum Engg · Algorithms
Industry Analyst
SCM · Data Center · Financial Research · VC
I have built Autonomous Agentic AI platforms, Data Mesh architectures, and Distributed systems that transform enterprise operations at scale. With proven delivery of $75M+ in savings through AI-driven optimization, PetaBytes scale data platforms and systems handling 500K+ TPS for millions of users, I bring hands-on expertise across the full stack - from Machine Learning, Distributed Systems, Data Engineering to Cloud-Native Infrastructure and Engineering Leadership.
Open to Consulting engagements, Technical collaboration, Mentorship, and Volunteer work with non-profits.
Let's discuss how I can help you build scalable systems, review your architecture, or accelerate your team's growth!
What I Can Help With
- →Architecture Reviews & Design
- →System Design & Technology Selection
- →Technical Due Diligence for M&A/Investments
- →AI/ML Strategy & Roadmap Planning
- →Data Platform & Data Mesh Architecture
- →AI Agents & RAG Systems
- →Data Engineering Pipelines & Platforms
- →Distributed Systems & Microservices
- →Cloud-Native Architecture (GCP, Azure)
- →High-Throughput & Low-Latency Systems
- →Engineering Leadership Coaching
- →Technical Career Guidance
- →System Design Interview Preparation
- →Code Review & Best Practices
- →Building High-Performance Teams
- →Non-Profit Technology Consulting
- →Educational Initiatives & Workshops
- →Open Source Project Contributions
- →Mentoring Underrepresented Groups in Tech
Portfolio
Scaled to 3 TB of curated data as the central engine for company-wide quarter-end reconciliation.
• Cut quarter-close cost from $140k → $8k.
• Certified for IBM ASCA & SOX compliance.
• Implemented access-identity security and zero-trust controls.
• Delivered end-to-end observability with Prometheus, Langfuse, and Grafana.
• Built data & AI governance frameworks for compliance and trust.
• Cut quarter-close cost from $140k → $8k.
• Certified for IBM ASCA & SOX compliance.
• Implemented access-identity security and zero-trust controls.
• Delivered end-to-end observability with Prometheus, Langfuse, and Grafana.
• Built data & AI governance frameworks for compliance and trust.
Data MeshData EngineeringIBM ASCASOXLangfuseObservability
Production multi-agent system powering a digital workforce across analytics, governance, reliability, and compliance.
• Dataverse Agent: NL-to-SQL, cutting analytics workload by 80%.
• Platform Review Agent: Automated governance, 30% cost reduction.
• Metadata Enrichment Agent: Fine-tuned PHI-3 model.
• DSR / GDPR Agent: Orchestrated scatter-gather for data-subject requests with full audit trail.
• SRE Agent: Dual knowledge-synthesis + worker agent — negotiated permissions, auto-remediated incidents, 45% MTTR reduction, 99.995% uptime.
• Dataverse Agent: NL-to-SQL, cutting analytics workload by 80%.
• Platform Review Agent: Automated governance, 30% cost reduction.
• Metadata Enrichment Agent: Fine-tuned PHI-3 model.
• DSR / GDPR Agent: Orchestrated scatter-gather for data-subject requests with full audit trail.
• SRE Agent: Dual knowledge-synthesis + worker agent — negotiated permissions, auto-remediated incidents, 45% MTTR reduction, 99.995% uptime.
Multi-Agent SystemsNL-to-SQLPHI-3OPAGovernance AutomationLLM Ops
Simulation harness that models humans, departments, and permissions to generate governed agentic clones — enabling safe agent development and policy testing.
• SRE agent pilot: Automated ticket resolution, peak-load scaling, onboarding validation, and self-improving knowledge base.
• Policy-Based Access: Dynamic attribute-based access control via OPA across micro-services.
• Ephemeral compute: Spawns research environments on-demand and destroys them post-session.
• SRE agent pilot: Automated ticket resolution, peak-load scaling, onboarding validation, and self-improving knowledge base.
• Policy-Based Access: Dynamic attribute-based access control via OPA across micro-services.
• Ephemeral compute: Spawns research environments on-demand and destroys them post-session.
Policy EngineOPAZero TrustAgentic SimulationNode.js
Led enterprise-wide Gen AI adoption across 129 associates, achieving 100% AI-assisted productivity.
• Drove adoption of Cursor, Claude CLI, and Gemini Enterprise.
• Ran rigorous evaluations of coding-review agents across Sonnet, Opus, Gemini Pro, and Kimi.
• Drove adoption of Cursor, Claude CLI, and Gemini Enterprise.
• Ran rigorous evaluations of coding-review agents across Sonnet, Opus, Gemini Pro, and Kimi.
Gen AILLM EvalsDeveloper Productivity
Built next-gen programmable dashboards and AI interaction hubs that replaced legacy BI tooling.
• Replaced Tableau with a SOX-compliant, dynamic visualization platform.
• Unlocked $2.3M in immediate savings while improving developer experience.
• Replaced Tableau with a SOX-compliant, dynamic visualization platform.
• Unlocked $2.3M in immediate savings while improving developer experience.
Developer ExperienceData VisualizationUX InterfacesSOX
Sweeping optimization engine for inventory placement across Walmart shipping nodes.
• Realized $75M of a projected $500M business impact.
• Enabled pixel-based demand forecasting to power next-gen AI fulfillment centers.
• Realized $75M of a projected $500M business impact.
• Enabled pixel-based demand forecasting to power next-gen AI fulfillment centers.
Optimization AlgorithmsDemand ForecastingSupply Chain
Multi-agent orchestration for placement analysis and approval workflows — collapsing cycle times by 92%.
• Reduced critical placement decisions from 48 hours → under 4 hours.
• Automated weekly analysis and cross-team approval pipelines.
• Reduced critical placement decisions from 48 hours → under 4 hours.
• Automated weekly analysis and cross-team approval pipelines.
Multi-Agent OrchestrationsGenAIDecision Automation
Achieved measurable cost and quality improvements through continuous platform optimization.
• 31% reduction in replenishment data lake operating costs via re-architecture and migration.
• 15% improvement in code quality and build metrics.
• Maintained 100% CVE compliance against zero-day exploits.
• 31% reduction in replenishment data lake operating costs via re-architecture and migration.
• 15% improvement in code quality and build metrics.
• Maintained 100% CVE compliance against zero-day exploits.
Cost OptimizationRe-architectureSecurity Compliance
Cross-functional platform bridging UX, Data, and QA — fully compliant with GDPR, POPIA, and CCPA.
• Reduced operational costs by 21.2%.
• Improved CSAT scores by 33.8% via an enhanced in-house conversational chatbot.
• Reduced operational costs by 21.2%.
• Improved CSAT scores by 33.8% via an enhanced in-house conversational chatbot.
Big DataConversational APIsChatbotGDPR/CCPA
Designed, built and delivered a Data Platform serving 4,000 US Walmart stores across 34 domain verticals with enterprise-grade reliability.
• Reduced schema onboarding from 1.5 weeks → 2 days.
• Built comprehensive Data Catalog & Quality framework for lineage and discovery.
• Achieved 99.99% uptime with >10TB daily ingestion under SOX & CCPA.
• Reduced schema onboarding from 1.5 weeks → 2 days.
• Built comprehensive Data Catalog & Quality framework for lineage and discovery.
• Achieved 99.99% uptime with >10TB daily ingestion under SOX & CCPA.
Data GovernanceBig DataSchema RegistrySOX/CCPAData Lineage
Next-generation Automated Inventory Management System consolidating fragmented platforms into a single source of truth.
• Consolidated 5 systems → 1 platform, reducing repos from 35 → 3.
• Delivered 40% cost savings and 500% throughput increase.
• Shipped a foundational ML library adopted by 6+ teams company-wide.
• Consolidated 5 systems → 1 platform, reducing repos from 35 → 3.
• Delivered 40% cost savings and 500% throughput increase.
• Shipped a foundational ML library adopted by 6+ teams company-wide.
Demand ForecastingSystem ConsolidationML LibrariesAnalytics Platform
Architected enterprise-grade Tier-0 inventory services operating at extreme scale and reliability.
• Sub-3ms latency at 13k+ TPS, processing 4M updates/hr across 22M items.
• Managed 1 TB Elasticsearch cluster for real-time search and forecasting on billions of records.
• Weekly >10TB delta loads improving forecasts by 35% for newly launched items.
• Sub-3ms latency at 13k+ TPS, processing 4M updates/hr across 22M items.
• Managed 1 TB Elasticsearch cluster for real-time search and forecasting on billions of records.
• Weekly >10TB delta loads improving forecasts by 35% for newly launched items.
High-Scale Distributed SystemsElasticsearchLow-LatencyData Ingestion
Patents
Skills
Multi-Agent SystemsLangGraphRAG (Retrieval-Augmented Generation)Fine-tuningVector StoresLLM Orchestration
Data MeshApache SparkAirflowSnowflakeDBTData Lakes
Google Cloud PlatformMicrosoft AzureKubernetesMicroservicesEvent-Driven ArchitectureDistributed Systems
OraclePostgreSQLCassandraElasticsearchRedisSQL & NoSQL
Supply Chain OptimizationRetail & eCommerceCompliance (SOX, GDPR, CCPA)MLOpsData Engineering
Experience
July 2025 - Present
Senior Engineering Leader – Data & AI — Red Hat
Leading the Data and AI Platform for India, managing a team of 28+ engineers across APAC, NA, and EMEA.
Driving the AI acceleration roadmap, scaling the Enterprise Data Mesh platform, and building the next-generation Agentic AI workforce.
Driving the AI acceleration roadmap, scaling the Enterprise Data Mesh platform, and building the next-generation Agentic AI workforce.
October 2012 - June 2025
Senior Architect / Senior Software Engineering Manager II — Walmart
Over a near 13-year tenure, progressed from SDE II to Architect and Senior Software Engineering Manager, leading large geo-distributed teams across some of Walmart's most critical and complex charters.
Directed the architecture, vision, and scale of massive enterprise initiatives including Inventory Management Systems, Real-Time Availability, Demand Forecasting, Store Systems Data Platforms, Centralized Comms Platform, and a flagship multi-million dollar Inventory Optimization System.
This trajectory built a deeply rounded profile encompassing the full technical stack—architecting high-scale Distributed Systems, modernizing PetaByte-scale Data architecture, deploying rigorous supply chain Optimization Systems, integrating foundational Machine Learning (ML), AI, and pioneering early orchestrations in Gen AI to radically elevate global fulfillment capabilities.
Directed the architecture, vision, and scale of massive enterprise initiatives including Inventory Management Systems, Real-Time Availability, Demand Forecasting, Store Systems Data Platforms, Centralized Comms Platform, and a flagship multi-million dollar Inventory Optimization System.
This trajectory built a deeply rounded profile encompassing the full technical stack—architecting high-scale Distributed Systems, modernizing PetaByte-scale Data architecture, deploying rigorous supply chain Optimization Systems, integrating foundational Machine Learning (ML), AI, and pioneering early orchestrations in Gen AI to radically elevate global fulfillment capabilities.
August 2010 - October 2012
Programmer - Senior Programmer — Best Buy (via Accenture BBTG)
Led critical retail system modernizations, including the overarching Sterling SSFS enterprise upgrade and the development of a unified mobile carrier activation platform.
Education
Georgia Institute of Technology
Master of Science (MS) in Machine Learning (In Progress)
Yeshwantrao Chavan College of Engineering (YCCE)
Bachelor of Engineering in Computer Technology