All Projects
Real problems, real solutions, real impact. Full Context → Approach → Impact for every project.
Managing enterprise-scale technology transformation with a lean team meant development velocity was the bottleneck. Traditional approaches — more developers, longer timelines — weren't viable. I needed to fundamentally change how we built and analyzed systems.
Designed and deployed an agentic AI architecture using the BMAD (Business-focused Multi-Agent Design) methodology. Built 58+ specialized AI agents, each with defined roles: codebase exploration, architecture planning, root cause analysis, database optimization, documentation generation, and code implementation. Created a token-optimized context system achieving 70-85% reduction in context overhead while maintaining full operational awareness.
Organizational knowledge was scattered across databases, documents, logs, and people's heads. When someone needed an answer, they asked around or dug through files. There was no unified way to query enterprise knowledge programmatically.
Architected a hybrid RAG (Retrieval-Augmented Generation) system combining vector search with graph-based knowledge representation. Built 45+ domain-specific collections across Qdrant for semantic search and Neo4j for relationship mapping. Designed ingestion pipelines using Docling for document processing, with Redis for caching and queue management.
The organization was spending heavily on telephony and call management through legacy providers. Collections calls, customer service, and sales outreach were entirely manual, with no transcription, no analytics, and no AI-assisted handling.
Designed an AI-powered voice agent platform integrating Google Gemini's multimodal capabilities with Exotel's telephony infrastructure. Built the end-to-end pipeline: inbound call handling, AI-powered conversation management, real-time transcription, and automated post-call analytics. Created N8N workflows for call routing, transcription processing, and CRM integration.
HR operations were fragmented across spreadsheets and disconnected tools — attendance tracking, leave management, payroll processing, and performance reviews all lived in different systems with no integration or single source of truth.
Architected an HRMS foundation using 30+ OCA (Odoo Community Association) modules, carefully selected and configured to cover the full employee lifecycle: recruitment, onboarding, attendance, holidays, contracts, payroll computation, and performance evaluation. Designed the module dependency graph to ensure clean upgrades and minimal customization debt.
The enterprise operated on 7+ disconnected systems: ERP, legacy operations platform, financial accounting, telephony, lead generation, marketplace, and messaging. Data was manually transferred between systems via WhatsApp messages, email, and spreadsheets — creating hundreds of daily data entry tasks, constant sync errors, and no single source of truth.
Designed a hub-and-spoke integration architecture with the ERP as the central data authority. Built API integrations, webhook listeners, and queue-based processing pipelines for each connected system. Implemented rate limiting, retry logic, and error aggregation to handle the realities of integrating systems with different reliability profiles.
A comprehensive security review uncovered critical vulnerabilities in a legacy operations platform: SQL injection exposure across hundreds of query patterns, SSL verification disabled in multiple locations, and no compliance framework for India's Digital Personal Data Protection Act 2025.
Conducted a systematic security audit covering code-level vulnerability analysis, infrastructure security assessment, and regulatory compliance gap analysis. Identified and cataloged every SQL injection vector, mapped data flow paths for personal data, and designed a multi-phase remediation strategy. Created a DPDP Act compliance roadmap covering consent management, data minimization, breach notification protocols, and cross-border transfer controls.
Business processes across departments were manual, repetitive, and error-prone. Lead capture, invoice reconciliation, compliance checks, and reporting all required human intervention at every step. The organization had no automation infrastructure.
Self-hosted N8N as the enterprise automation platform — a deliberate choice over cloud alternatives to maintain data sovereignty and reduce costs. Designed and deployed 20+ production workflows spanning AR/finance automation, sales/marketing lead pipelines, IT infrastructure monitoring, and compliance validation. Each workflow was built with error handling, retry logic, and monitoring.
Standard Operating Procedures existed as static documents that nobody read — buried in shared drives, outdated within weeks of creation, and disconnected from the systems they described. When processes changed, documentation lagged by months.
Designed an interactive HTML5 SOP playbook system with rich visual elements: architectural diagrams, decision trees, process flow visualizations, and step-by-step interactive guides. Built each SOP as a self-contained web application with search, navigation, and visual process mapping. Created 96+ architectural diagrams to make complex multi-system processes comprehensible at a glance.