SKGS
INITIALIZING FLIGHT SYSTEMS
00%
Experience Focus Areas Open Source Writing Contact
BLR · BENGALURU, KARNATAKA, INDIA
Airline Domain → Enterprise AI

SHASHI
KANTH

G S.

An architect working at the intersection of deep airline domain knowledge and AI, with business acumen for judging what is worth solving and delivering it at enterprise grade, at scale, and governed.

Author & maintainer of allsrc.dev · View on GitHub
View Experience
Scroll
0+ Years in Aviation Tech
0 IPs
0+ Airlines Delivered For
Airline Domain◆ Offer & Order◆ NDC Architecture◆ Agentic AI◆ AIOps◆ Platform Engineering◆ Enterprise Scale◆ AI Governance◆
Shashi Kanth G S
Solutions Architect · Airline Domain & Enterprise AI

Where Airline Domain
Meets Enterprise AI.

I find where the cost or risk actually lives, prove the idea fast through a POC, MLP, or MVP, weigh the risk of acting against the risk of not acting, and then shape architecture with resilience, scale, and governance designed in from the start.

Airlines & programs delivered for
Hahn Air Iberia Lufthansa Group British Airways TravelSky (GDS) Saudia UK Ministry of Defence

A Career Built
in Aviation Tech.

Current
Amadeus (Amadeus Software Labs India Pvt Ltd)
Solutions Architect
Current

Leading cloud transformation and middleware modernization initiatives for airline platforms, with emphasis on distribution APIs, enterprise integration, operational architecture, and data engineering across ETL pipelines and bronze/silver layer design.

Cloud TransformationMiddleware ModernizationDistribution APIsEnterprise IntegrationData EngineeringETL PipelinesBronze & Silver Layer Design
Previous
Unisys
Consultant → Tech Lead → Technical Architect

Progressed across architecture and technical leadership roles while shaping airline PSS, offer and order management systems, AWS cloud microservices, retailing integrations, deployment patterns, and recovery strategies.

Airline PSSOffer & Order ManagementAWS CloudMicroservicesRetailing IntegrationsDeployment PatternsRecovery Strategies
Early Career
NTT DATA
Trainee → Developer → Senior Software Engineer

Started in airline passenger systems delivery, growing quickly into client-facing engineering and strategic cutover work.

Airline PSSClient DeliveryCutover Strategy

Where Deep Work
Gets Done.

01

Application & Platform Modernization

Re-architecting legacy airline platforms into cloud-native, resilient systems — modernizing middleware, decomposing monoliths, and clearing the path for faster, safer delivery.

Cloud MigrationMiddlewareLegacy DecompositionResilience
02

Airline Retailing

Architecture across the retailing value chain — merchandising, personalized offers, ancillaries, and the booking experience that turns a fare into a relationship.

RetailingMerchandisingAncillariesBooking Experience
03

Passenger Service Systems (PSS)

Deep, hands-on architecture across reservations, inventory, and departure control — the systems of record that keep an airline flying.

ReservationsInventoryDCSPassenger Systems
04

Offer & Order Management

NDC-native offer construction and order lifecycle management — decoupling the offer from the fare, and the order from the ticket.

NDCOffer ManagementOrder ManagementDistribution APIs
08

Developer Tools & Platform Enablement

Open-source tooling that closes real gaps in the agent ecosystem — protocol bridges, governance harnesses, and reference implementations other engineers actually reuse.

A2AMCPOpen SourceTooling
09

Homelab & Private Platform Operations

A proving ground I maintain passionately, end to end, across a fleet of self-hosted servers — API gateways, WAF, AI-driven observability, automation, security, firewalls, hosting, and Kubernetes.

KubernetesAPI GatewayWAFObservabilityAutomationSecurity

Building in
the Open.

Visit GitHub Profile

Problem AI coding agents such as Claude Code, GitHub Copilot, Antigravity, and OpenCode do not speak the Agent2Agent (A2A) protocol out of the box, so they cannot be discovered or orchestrated alongside other agents.

A2A ProtocolAny AgentOrchestration
GitHub Docs

Problem A2A agents and MCP tool servers are separate ecosystems, so an A2A agent's skills are not reusable as MCP tools without custom glue code.

A2AMCPSkill Reuse
GitHub Docs

Problem A2A agents are usually tested as raw protocol endpoints, which makes it hard to see how they behave as real, stateful applications.

A2A ProtocolTestingStateful Agents
GitHub Docs

Problem Most agent frameworks address how an agent thinks, but leave the harness and governance layer — what the agent is allowed to do — unaddressed and untested.

Agent GovernanceHarnessTesting
GitHub

Problem Giving an AI assistant direct SSH access to infrastructure means either exposing raw credentials, or blocking it from operational tasks entirely.

MCPSSHSecurityInfrastructure
GitHub

Problem Teams adopting A2A need to see the protocol working end-to-end inside a real agent framework, not just read a spec.

A2A ProtocolReference ImplTravel Domain
GitHub

Technical Writing.

allsrc.dev

Your Framework Choice Matters Less Than You Think

The same governance patterns, unmodified, on LangGraph and Microsoft Agent Framework — with parity tests and the footguns found in each.

Twelve Governance Patterns On One Agent: Do They Compose?

Individual patterns are easy to believe in. Running twelve together on one agent found a real gap none of their own tests could have caught.

"Which Agents Do We Have In Production?" Is A Bad Question To Fail

A spreadsheet of agents drifts the day after you write it. Make the profile enforced: owners, declared tools, expiry, and privilege-drift detection.

Open to conversations on

Architecture.
Agentic AI.
Airline Domain.

Writing
allsrc.dev
LinkedIn
shashikanthgs
GitHub
shashikanth-gs