---

🏛️ The Vision: Digital Armor for an AI-First Nation

I am an AI/ML Engineer & Full-Stack Architect based in India 🇮🇳, dedicated to building the essential trust layer that AI systems in Bharat require to move from prototype to production.

In a rapidly digitizing economy, AI governance isn't just a hurdle — it's the foundation of trust. My work focuses on creating Digital Armor for enterprise AI, ensuring that as systems proliferate across Fintech, Healthcare, HR, Education, and Public Infrastructure, they remain secure, fair, and perfectly aligned with the digital laws of the land.

🛡️ KavachX: The Enforcement Layer

My flagship project — KavachX — is a production-ready AI governance engine designed to serve as the trust and enforcement layer between AI models and the end users they serve. Think of it as a compliance firewall for AI — every inference is scored, audited, and enforced against a live policy context before it ever reaches a user.

What KavachX does:

• 🔍 Real-time risk scoring — every AI inference passes through a composite risk engine that evaluates content safety, demographic fairness, data privacy, and adversarial intent simultaneously.
• 🧬 Fine-tuned ML safety classifiers — custom-trained BERT-based models (not just rule-based filters) that learn from domain-specific data and improve over time.
• 🔒 Tiered enforcement gates — soft reviews, hard blocks, and escalation paths based on risk thresholds; no binary on/off switch.
• 📡 Immutable audit trail — every enforcement decision — including blocks — is logged, timestamped, and surfaced via an executive compliance dashboard in real time.
• 🧩 Plug-and-play middleware — wrap any LLM or ML API behind a single KavachX key; no model retraining required.
• 🕵️ Synthetic media forensics — early-stage detection of AI-generated media abuse through the integrated media shield module.
• 🌐 Browser-layer governance — a Chrome extension intercepts AI interactions in the browser itself, extending enforcement beyond the server.

Industries KavachX is designed for: Fintech · HealthTech · EdTech · HRTech · LegalTech · E-commerce · Government & Public Sector

KavachX follows all applicable AI-related and digital governance laws operative in Bharat as its compliance baseline — with deep focus on the Digital Personal Data Protection Act and general AI safety standards.

---

Beyond KavachX:

• 💪 Built Mpower Fitness — a full-stack multi-role SaaS with real-time Socket.IO chat, UPI payments, and trainer analytics.
• 🤖 Built PO CoPilot — a live Streamlit AI assistant for SAP-aligned procurement workflows.
• 🛒 Building VyaparGPT — an LLM-powered business intelligence assistant for Indian SME traders.
• 🤝 Open to collaborations on AI safety, responsible ML, and enterprise B2B SaaS.

---

🚀 Featured Projects

🛡️ KavachX — AI Governance Engine The enforcement layer for production AI A real-time governance platform acting as the middle-layer for AI interactions. Implements composite risk scoring and immutable audit logging. Highlights: ⚡ Real-time inference risk scoring 🔒 Bharat-native compliance engine 🧬 ML fine-tuned safety classifiers 🧩 Browser-layer AI interception 🐳 Dockerized • Nginx-proxied

💪 Mpower Fitness Platform Full-stack SaaS for the fitness economy A multi-role fitness SaaS with real-time trainer-client management, analytics, and indigenous payment integrations. Highlights: 💬 Real-time trainer↔client chat via Socket.IO 📊 Complex analytics & streak tracking 💳 UPI deep-link payment ecosystem 🏋️ Full workout library with timer 🗄️ PostgreSQL with Sequelize ORM

🤖 PO CoPilot — AI Procurement Assistant AI-powered Purchase Order validation & change impact analysis A Streamlit-based AI assistant for procurement teams to validate POs, analyze change impacts, and generate approval summaries — aligned to SAP BTP architecture. Highlights: 🔍 AI-powered PO validation with change impact analysis 📋 Automated approval summary generation ☁️ SAP BTP-aligned architecture 🚀 Live Demo

🛒 VyaparGPT — Business Intelligence Assistant BI for the Bharat SME market An LLM-powered assistant enabling Indian traders to query business data using natural language in Hindi and English. Highlights: 🗣️ Bilingual business intelligence 📈 AI-driven sales & inventory insights 🇮🇳 Tailored for SME traders

---

🛠️ Tech Stack

Languages

AI / ML

Backend

Frontend

Data & Infra

---

📊 GitHub Stats

 

---

🏆 GitHub Achievements

---

💬 What I'm Working On

class TheIndicSentinel:
    def __init__(self):
        self.name        = "The Indic Sentinel"
        self.focus       = "AI Governance & Responsible ML"
        self.location    = "India 🇮🇳"
        self.flagship    = "KavachX — Production AI Compliance Engine"
        self.stack       = ["Python", "FastAPI", "React", "PyTorch", "Docker"]
        self.compliance  = ["DPDP Act 2023", "All applicable AI & digital laws in Bharat"]
        self.open_to     = ["Collaborations", "AI Safety Research", "Enterprise SaaS"]

    def current_sprint(self):
        return [
            "🧬 Fine-tuning BERT-based safety classifiers",
            "🔒 Hardening DPDP policy enforcement engine",
            "📡 Real-time governance dashboard via WebSockets",
            "🧠 Expanding synthetic media forensics module",
        ]

---

"In the age of AI, the question is not just what machines can do — but what they should do."

⭐ Star my repos if you find them useful!