Sovereign AI Advocate | Global FinTaxTech Leader | 20+ Years Enterprise Experience
"I build AI that runs enterprise operations across 54 countries, not only slide decks."
📍 India | 📧 Akhilesh.digital@Outlook.com | 🔗 LinkedIn | 📦 GitHub
- Hands-on AI Builder: I personally architect production AI systems, not just lead strategy
- Enterprise Scale: Systems I've architected support a Multi-million $+ transaction footprint across 54 jurisdictions
- CFO-Level Impact: My AI systems save millions annually, recognized with CFO 100 Award (2023-2025)
- Sovereign AI Advocate: In-region, sovereign LLM deployment as a governance principle
- Global Delivery: Directed onsite programs across USA, UK, Australia, Singapore, Canada, delivery focused
The toolset actually running behind enterprise systems at LDC, Azure-native, chosen for data residency and enterprise governance.
| Layer | Tools |
|---|---|
| Orchestration & Agents | Semantic Kernel · Azure AI Foundry · AutoGen (single- & multi-agent) |
| LLMs | Azure OpenAI Service, GPT-4o, o1 (in-region / sovereign deployment) |
| Retrieval | Azure AI Search (vector index) · Azure OpenAI text-embedding-3-large |
| Evaluation & Tracing | Prompt Flow |
| Data & Analytics | MS Fabric · Databricks · Snowflake · SQL (MS SQL Server, Oracle PL/SQL) |
| ERP Integration | SAP S/4HANA · ONESOURCE · OneStream · D365 F&O |
These are the tools behind the RAG compliance pipeline, invoice anomaly detection model, and legal document automation engine described on my CV, same systems, same names, no translation needed between the two.
A lighter-weight stack I use for demos and personal experimentation, not the enterprise toolset above, and not presented as such.
30+ agentic AI use cases built across this Test Bed, a mix of single-agent workflows and multi-agent orchestrations, several of which were piloted or rolled out to production at LDC. The 4 repos below are representative examples, not the full set.
| Agent | Purpose | Stack |
|---|---|---|
| 🔍 last30days-skill | Multi-source research aggregation (Reddit, X, YouTube, HN) | LangGraph orchestration |
| 🧠 llm-council-skill2 | Multi-LLM deliberation for hallucination reduction | Claude 3.5 Sonnet, Gemini 1.5 Pro, Phi-3 |
| ⚡ gstack | Prompt methodology experiments | Claude Code CLI |
| 🛡️ AgentFS | Auditable AI filesystem, verifiable audit trails for agent decisions | Python + SQLite |
RAG pipeline (Test Bed): FAISS + Sentence-Transformers (open-source embeddings), the lightweight counterpart to the Azure AI Search + Azure OpenAI embeddings used in production.
Demo-scope only, built to validate patterns and technical feasibility before proposing them for enterprise use, not run on enterprise infrastructure.
A few samples from the 50+ agentic AI use cases in my Test Bed, several of these were piloted or rolled out to production at LDC. Sorted by value-creation potential (my own qualitative assessment of business impact, not a per-agent production or ROI claim).
| Agent | Category | Problem it addresses | Value-Creation Potential |
|---|---|---|---|
| Financial Reconciliation Investigation | Finance | Tracing and explaining mismatches between ledgers/accounts that would otherwise need manual line-by-line investigation | High |
| Journal-Entry Audit | Finance | Reviewing journal entries for anomalies or policy violations before period close | High |
| Regulatory Report-Drafting | Compliance | Assembling first-draft regulatory filings from underlying transaction data | High |
| Financial-Crime False-Positive Reduction | Compliance | Cutting the volume of flagged transactions that turn out to be legitimate, so investigators focus on real risk | High |
| Procure-to-Pay | Procurement | Automating the invoice-to-payment cycle end to end | High |
| Invoice Exception & 3-Way-Match | Procurement | Catching mismatches between purchase order, goods receipt, and invoice before payment | High |
| Supply-Chain Forecasting & Transshipment | Supply Chain | Forecasting demand and flagging optimal transshipment points across a logistics network | High |
| Spec-to-Build Developer | ERP | Turning a written functional spec into a first-pass code/configuration build | Medium |
| Release/Regression-Testing | Delivery | Running and triaging regression test suites ahead of a release | Medium |
| Change-Management/Change-Ticket | Delivery | Drafting and routing change tickets through approval workflows | Medium |
| Knowledge-Capture | Operations | Turning project artifacts/conversations into reusable documented knowledge | Medium |
| Alert Correlation & Noise-Reduction | Operations | Grouping related monitoring alerts into one incident instead of many duplicate pages | Medium |
| Ticket Triage & Auto-Resolution | Operations | Classifying incoming support tickets and auto-resolving the common/repetitive ones | Medium |
| Recruitment Screening | HR | Screening incoming candidate applications against role requirements | Medium |
| HR Onboarding & Associate-Allocation | HR | Coordinating new-hire onboarding tasks and initial team/project allocation | Medium |
| Access-Provisioning Request | Workforce | Handling system-access requests and routing them for approval | Medium |
| Carrier-Rate Comparison | Logistics | Comparing freight carrier rates/routes to surface the lowest-cost option | Medium |
| Shipment Tracking & Exception | Logistics | Monitoring shipments in transit and flagging exceptions (delays, damage, misroutes) | Medium |
(IT & Infrastructure and Commercial & Customer categories exist in the broader portfolio but are left off this table for now.)
| System | Impact | Scale |
|---|---|---|
| Tax Resolution Agent (RAG) | 65% research time reduction | 54 countries |
| Invoice Fraud Detection (Naive Bayes) | 95% accuracy in production | Global commodity transactions |
| VAT Compliance Automation | 80%+ straight-through processing | Europe & Brazil |
| Cloud FinOps Optimization | $2M annual savings | Azure / AWS / GCP |
| AI Governance Framework | ISO 42001 compliant | 54 jurisdictions |
- CFO 100 Award (2023-2025): Digital Transformation Excellence at Louis Dreyfus Company
- Most Valued Person - MVP (2024): Exceptional leadership at LDC India
- KPMG Encore Team & Kudos (2016-2017): Digital Innovation & Enterprise Solutioning
- Deputy CEO Appreciation (2016): Strategic transformation impact at KPMG
- NIIT Award of Excellence (2001): 300%+ productivity gain via automation
- Global Scale: Deployed systems across 54 countries, managing $10M+ budgets
- Team Leadership: Built 120+ FTE teams; established GCC from greenfield (50+ engineers in 90 days)
- CXO Relationships: Trusted advisor to 200+ CXOs; reported directly to Global CFO at LDC
- Certifications: ISO 42001 Lead Implementer | CAISO | CAIO | AB 731 AI Transformation Leader (2026) | PMP | PRINCE2 | Six Sigma Green Belt
- (Prior, expired: Microsoft Azure Certification, Enterprise SQL Certification, early cloud/data credentials)
I'm available for Senior Leadership Roles in:
- AI Transformation Director / VP
- Chief AI Officer / Head of AI
- Global Digital Transformation Lead
- GCC / Capability Centre Head
Available to join within 10/15 days.
📧 Akhilesh.digital@Outlook.com
🔗 LinkedIn | 📦 GitHub
"I don't build slide decks. I build systems that run enterprises."