Agentic AI Architect. I architect, build, and run cloud-native distributed systems end-to-end — from system design through cross-team coordination to production operations. Client-facing tech lead for 15+ enterprise engagements. Most recently Multi-Agent AI on Google ADK, the Microsoft Agent Framework, MCP, and Vertex AI Agent Engine.
I'm an Agentic AI Architect with 7+ years architecting, building, and running production cloud-native systems across FinTech, Open Banking, AdTech, and Telecom on Google Cloud, AWS, and Azure. I own the full lifecycle — from system design and architecture through team leadership and cross-functional coordination to production deployment and client-facing delivery.
Most recently, I was Agentic AI Architect at Searce (Google Cloud Premier Partner), where I led the BigQuery FinOps engagement for Tata Group — driving a 57% data warehouse cost reduction.
I architect and ship production systems at the intersection of Multi-Agent AI (Google ADK · Microsoft Agent Framework · MCP · Vertex AI Agent Engine · Amazon Bedrock Agents · LangChain · CrewAI) and production backend engineering (Go, Python, TypeScript, gRPC, GraphQL, Kubernetes). I'm a core contributor to Google's open-source Spanner Migration Tool (40–60% post-migration query gains), and I designed and shipped two production-grade open-source AI platforms — Genie (multi-agent financial assistant on the Microsoft Agent Framework) and Bodh (medical AI diagnostic panel extending Microsoft's MAI-DxO).
I lead from the front — architecting systems, building them hands-on, coordinating across engineering, product, and client teams, and running them in production. Whether it's a 30K TPS financial transaction engine or a multi-agent AI platform for regulated industries, I take end-to-end ownership: system design, implementation, cross-team alignment, deployment, and ongoing operations.
Systems I architected, built, and ran in production for enterprise and consumer clients — delivering significant FinOps savings, 30K+ TPS transaction engines, 1M+ user platforms, and 100K+ concurrent voting systems. End-to-end ownership from design through client delivery.
Architected and delivered Tata Group's BigQuery FinOps transformation — led cross-team coordination with Tata's data engineering teams, redesigned query patterns and MERGE operations, transitioned to capacity-based slot reservations. Delivered 57% data warehouse cost reduction.
Architected and led a 10-engineer team building a Kubernetes transaction platform processing 30K+ TPS for telecom/FinTech partner integrations. Owned the full stack from architecture through client delivery — PCI-aligned design, idempotent processing, error-code orchestration with exponential backoff and DLQ resilience.
Zero-trust voting and analytics platform for the AFL Brownlow Medal — 100K+ votes, 10K+ concurrent users on Cloud Run during live broadcasts, secured with Cloud KMS + Security Command Center.
Led a remote-first international team of 4 engineers, architecting and running the backend for a social platform serving 1M+ users. Reduced API response latency 47% and infrastructure costs 40% via protobuf contracts and service consolidation. Owned architecture, cross-team coordination, and production operations; 80%+ automated test coverage.
Architected and delivered a regulatory-grade Open Banking portal for multi-jurisdiction Middle East markets. Built consent management, multi-region data residency, and RAG-based semantic search on AlloyDB AI. Led cross-team coordination with banking partners — 37% latency reduction.
Upstream contributions to Microsoft's official Agent Framework SDK, plus production-grade open-source AI systems I designed, built, and shipped as reference implementations on that framework. Each demonstrates end-to-end engineering: architecture, implementation, security, governance, observability, and production deployment.
Merged contributions to Microsoft's official Go SDK for the Agent Framework — runtime-correctness and safety fixes surfaced by running real multi-agent systems in production: a nil-dereference SIGSEGV in the tool-approval path, an empty-choices crash on the Chat Completions response, and a dead-code handler-collision guard. 41 pull requests to the SDK — 12 merged into main, 29 open.
Go-based multi-agent financial assistant on the Microsoft Agent Framework — 15 role-specialized agents (supervisor, analyzer, forecaster, anomaly_detector, recommender, llm_auditor, etc.), JWT + RBAC, AES-256-GCM envelope encryption, full OpenTelemetry tracing.
Original open-source medical AI platform on the Microsoft Agent Framework — virtual physician panel inspired by Microsoft's MAI-DxO + SD-Bench. Role-specialized agents (intake, questioner, test_planner, cost_guardian, diagnostician, reasoning_verifier); FHIR R4 + HL7 v2 aware; cost-aware diagnostic budget enforcement.
Core contributor to Google's open-source Spanner migration tool — architected and built backend APIs for the Intelligent Schema Assistant, designed CDC pipelines (Datastream · Pub/Sub · Dataflow), and improved post-migration query performance by 40–60% through PK design, indexing, and table interleaving.
In 2025–2026, I architected and executed a strategic migration of 18+ production agents from Google's Application Development Kit (ADK) to the Microsoft Agent Framework — designing portable agent architectures, abstracting provider dependencies, and implementing governance frameworks end-to-end. Both Genie (multi-agent financial advisor) and Bodh (medical AI platform) are Microsoft Agent Framework reference implementations I built and run. Published 8 detailed technical case studies covering orchestration patterns, provider abstraction, tool governance, state management, and production deployment.
The philosophy, architectural clarity, and vendor lock-in concerns that drove the migration. Provider abstraction as the foundation for portable agents.
How orchestration shifts from implicit callbacks (ADK) to explicit executor loops (Microsoft Agent Framework). Why this matters for observability, auditability, and control.
Porting session state from ADK's opaque state dict to the Microsoft Agent Framework's explicit AgentThread. Token budgeting, long-term memory, and conversation audit trails.
Porting tools and adding governance layers (audit, approval, policy). DLP enforcement with OPA. Multi-step tool composition and error handling.
Zero-config local (Ollama), cost-optimized (OpenAI), and production (Azure Foundry) provider setup. Hybrid multi-provider orchestration.
Porting callbacks to middleware. Composable decorators for audit logging, retry logic, token enforcement, and OpenTelemetry integration.
Cloud Run deployments, agent-to-agent communication, load balancing, graceful shutdown, and production observability integration.
What worked, what was hard, and what we'd do differently. Real numbers: agent count, timeline, bugs found, provider swaps, governance policies built.
A few of 13 LinkedIn recommendations from managers, peers, and enterprise clients.
“Pratik led a team of four engineers, balancing hands-on technical depth with genuine leadership. Any team would be fortunate to have him, and I recommend him without hesitation.”
“A strategic cloud cost optimization program that significantly reduced overall BigQuery warehouse costs. What stands out about Pratik is his ability to work as a true partner.”
“If there’s one word that defines him, it’s intellectually relentless. Someone who will constantly push technical boundaries and raise the bar for engineering standards.”
Technical analysis of migrating 18+ multi-agent systems in 90 days. Explores architectural patterns, provider abstraction strategies, and key lessons from production deployments.
How we achieved 57% BigQuery cost reduction — slot reservation transitions, MERGE anti-pattern elimination, and capacity-based pricing strategies for enterprise data warehouses.
Architecture of a telecom/FinTech transaction engine processing 30K+ TPS — idempotent processing, PCI-aligned design, error-code orchestration with exponential backoff and DLQ resilience.
View All Articles (105+) Technical Articles Subscribe via RSS
In-depth technical articles on architecting and running multi-agent AI, security, governance, and the production realities of building reliable AI systems in regulated environments. Every post cites working code in github.com/c2siorg/genie.
Every Professional Cloud Security Engineer exam bullet, mapped to a file path in an RBI FREE-AI aligned Go platform. Where the implementation matches, where the analog substitutes, and where the honest gaps are.
Request → N-eyes approve → window-of-time → automatic expiry, with every transition written to a hash-chained audit log. The package that closes Gap #1 from the PCSE map.
The mental model that says no two adjacent layers share a single point of failure for the same class of attack. From TLS to OTel, the eleven layers a customer request crosses before an answer comes back.
Twelve months of running multi-agent AI in a regulated context. SLIs that matter, the incident runbook, drift detection, continuous adversarial testing, secret rotation, compliance posture as code.
The long-form security narrative for a multi-agent financial assistant — authentication, authorisation, tenant isolation, dual-identity audit, envelope encryption, hash-chained logs, governance, red team, BCP.
Board policy as a YAML file the risk team owns. Annexure VI as a database query. Every governance recommendation rendered as a file path in a Go repository.
Stdlib over libraries, single binary over framework, fail-closed defaults over forgiveness. The boring-on-purpose case for choosing Go to ship a multi-agent system into a regulated environment.
Every one of the 26 RBI FREE-AI recommendations, mapped to a specific file in a working multi-agent platform. What's ✅ done, what's 🟡 partial, what's ⚪ honest gap.
Go · Multi-cloud
Go · Compiler · Blockchain
I architect, build, and run production systems end-to-end. Open to Agentic AI Architect and Staff Engineer roles in FinTech, Open Banking, and AI infrastructure. I also provide strategic consulting through direct engagement.
Based in Pune, India · Available for Remote Opportunities · Operating Hours: IST (UTC+5:30) with flexible availability across US, EMEA, and APAC time zones.
Professional Recognition & Impact
Google Summer of Code Mentor (2017–2026, 40+ students mentored) · Google Cloud Next 2022 Conference Speaker · Core Contributor to Google Cloud Platform's Spanner Migration Tool