Amrik Singh Khalsa

    Also known as Amrik | Amrik Singh

    AI Agent Systems & Software Architect

    Hi, I'm Amrik, a Bangalore-based software architect building AI agents, multi-agent frameworks, and the MCP, identity and security infrastructure that makes them safe to run in production. Before that, fintech platforms at scale.

    0+
    AI Agents & MCP Servers Shipped
    0
    Agent Platform PoCs, End-to-End
    0+
    PRs Merged on AI Platform & Security
    0+
    Architecture & Research Docs Published
    "Measure before you claim. Read the source, not the summary. Retract in public. Most of the systems I've built exist because a number was being quoted that nobody had verified."

    About Amrik Singh Khalsa

    Amrik (also known as Amrik Singh) is a software architect building AI agents and the frameworks, protocols and security gates that let them run in production, with a full-stack engineering foundation underneath and a bike trip never far away.

    Amrik Singh Khalsa - Professional Portrait
    AI Agent Architect
    Agent Security & Governance
    Open Source Contributor
    Startup Builder
    Full-Stack Architect
    Top 1% Problem Solver

    I'm Amrik Singh Khalsa, a software architect working on AI agent systems. Most of what I build these days is agent infrastructure: multi-agent pipelines that root-cause production problems with cited evidence, MCP servers, agent-to-agent protocol runtimes, and the security gates that make executable agent skills safe to distribute across an engineering organisation.

    The habit that runs through all of it is verification. A disproportionate share of my work exists because a number was being quoted that nobody had checked, so I build the instrumentation, read the source instead of the summary, and publish the correction when the answer turns out to be different. Including when the claim I'm correcting is my own.

    I contribute upstream to open source too. Most recently I shipped a merged fix to Warp, the Rust terminal, so executable shell scripts opened from a file:// URL run in the terminal instead of landing in the editor. Thirteen unit tests came with it, because a behaviour change gated on a file permission bit is exactly the kind of thing that needs them.

    Before this, I was a Founding Engineer at Incentiv, where I built a platform that enabled $5M+ in ESOP liquidity, onboarding 30+ startups and 200+ investors in six months. My journey started with founding PRAYATNA, a non-profit where I mentored underprivileged students, teaching me that technology's true power lies in empowering others. At Augmento Labs I led frontend teams, cut infrastructure costs, and built reusable systems teams actually wanted to use.

    When I'm not crafting scalable architectures, you'll find me planning my next bike trip, searching for that perfect sunset spot on the beach, or discovering a hidden café during late-night coding sessions. I'm passionate about cricket, love working from mountain retreats, and believe that the best code comes from a mind that's been refreshed by adventure.

    Core Values

    Evidence Over Assertion

    Measure it before you claim it, and read the source rather than the summary

    Honest Degradation

    A system that says what it couldn't reach beats one that confidently invents

    Adventure-Driven Innovation

    Best ideas come from fresh perspectives and new experiences

    Resilience & Adaptability

    Like biking uphill, persistence leads to amazing views

    Meaningful Impact

    Building solutions that empower people and create real change

    Continuous Learning

    Every sunset, every trip, every challenge teaches something new

    Experience Timeline

    My journey through technology leadership, building scalable systems and driving innovation across diverse industries.

    Current
    Oct 2025 - Present

    Senior Software Engineer

    Okta (Auth0)

    Building AI agent infrastructure and the security, identity and governance layers around it: multi-agent frameworks, MCP servers, agent-to-agent protocols, and the CI gates that make executable agent skills safe to distribute across an engineering organisation.

    Own the security gates for an org-wide AI agent skill marketplace: scanner, static-analysis hook and duplicate gate, all merged, with OIDC-federated cloud credentials in CI
    Architected a multi-agent resolution pipeline that root-causes production escalations with cited evidence, promoted from a personal skill to an org-wide plugin
    Shipped six agent-platform proofs of concept end to end, covering A2A runtimes, MCP servers, agent memory with retrieval grounding, orchestration and observability, each with a real test suite
    Published deep research on agent identity and delegated authorisation, verified against server source rather than internal docs, plus an agent migration design onto managed cloud runtimes
    Built AI-powered support automation with LLM and RAG over internal documentation
    Claude / Agent SDKs
    MCP
    A2A
    AWS Bedrock
    TypeScript
    Python
    Go
    Terraform
    Previous
    Apr 2024 - Jun 2025

    Founding Engineer

    Incentiv

    Leading technical innovation in fintech by architecting scalable, high-performance systems that process millions in transactions while maintaining enterprise-grade security.

    Architected scalable platform processing $5M+ in transactions
    Designed distributed systems handling 10,000+ monthly operations
    Delivered 25% performance improvement through system optimization
    NestJS
    Next.js
    PostgreSQL
    Redis
    Previous
    Jun 2021 - Apr 2024

    Software Engineer → Frontend Lead

    Augmento Labs

    Led technical teams through critical growth phases while architecting shared infrastructure and driving innovation across multiple product lines.

    Led development teams across 6+ complex frontend applications
    Reduced infrastructure costs by 30% through optimization strategies
    Architected migration solutions delivering 44% performance improvements
    React
    TypeScript
    Azure
    .NET

    Featured Projects

    Recent work on AI agents, agent security and platform architecture, alongside the fintech systems that came before it. Internal names, identifiers and findings are generalised.

    AI Agents

    Multi-Agent Resolution Pipeline for Production Support

    An orchestrator and specialist agent fleet that root-causes escalations with cited evidence

    Problem

    A single customer escalation cost an engineer a full day of manually correlating issue trackers, internal docs, source history and chat threads, and the reasoning behind each answer was never reproducible.

    Solution

    Designed an orchestrator, analyst and synthesizer agent fleet. Four specialist subagents run in parallel over separate evidence sources, each emitting a confidence-rated JSON artifact, then a synthesizer composes the final resolution. All network I/O and persistence stay with the orchestrator, so a subagent can never hallucinate that it wrote a file.

    Impact

    Promoted from a personal skill to an org-wide plugin. Root-caused a dozen escalations end to end, including a rate-limiting defect whose own test had been updated to assert the bug, which is why CI never flagged it. When a source is unreachable the pipeline says so on the face of the output instead of quietly filling the gap.

    Multi-Agent Orchestration
    MCP
    Agent SDKs
    Node.js
    Python
    Evidence Grading
    Multi-Agent Resolution Pipeline for Production Support
    AI Security

    Security Gates for an Org-Wide AI Skill Marketplace

    Making executable agent skills safe to distribute across an engineering organisation

    Problem

    An internal marketplace distributes executable AI agent skills to every engineering team. Anyone could publish, nothing scanned what shipped, and the risk headline in circulation had never been derived from the raw findings.

    Solution

    Owned and shipped three merged CI gates: an LLM-backed skill security scanner, a static-analysis pre-commit hook, and a duplicate-skill gate. Added the Terraform module and OIDC-federated IAM roles that give the scanner short-lived model credentials in CI with no long-lived secrets.

    Impact

    Scanned the full skill catalogue and published the results as a structured per-plugin report set. Re-derived the circulated risk headline and corrected it in public, because it described none of the severity everyone had assumed. Also caught a silent fail-open in report naming that was quietly dropping skills from the gate.

    Terraform
    GitHub OIDC
    AWS Bedrock
    Static Analysis
    LangGraph
    CI/CD
    Security Gates for an Org-Wide AI Skill Marketplace
    Platform Architecture

    Agent-to-Agent Runtime, Identity & Protocol Architecture

    Designing how autonomous agents authenticate, delegate work, and talk to each other

    Problem

    Agent platform plans rested on assumptions nobody had verified in code: which inbound auth modes can coexist on a runtime, whether an in-network token issuer was even reachable, and whether agent context usage was measurable at all.

    Solution

    Built an A2A protocol runtime alongside the existing HTTP agent runtime, an MCP server exposing read-only service-catalog tools, an agent memory and retrieval-grounded knowledge layer on managed cloud primitives, and a context-utilization metric wired from the agent SDK through the event contract to the UI.

    Impact

    1,200+ tests green across the A2A build. Found that the platform's own OIDC issuer already satisfied a requirement the team thought needed a new dependency, and surfaced a missing cloud-provider field that would have silently blocked the design before anyone built on it. Proved context utilization was computable today, removing a dependency on a deferred metrics programme.

    A2A Protocol
    MCP
    Bedrock AgentCore
    TypeScript
    Go
    Terraform
    OIDC / JWT
    Agent-to-Agent Runtime, Identity & Protocol Architecture
    Open Source

    Upstream Fix in Warp, the Rust Terminal

    Executable shell scripts opened from a file:// URL now run instead of opening in the editor

    Problem

    Warp's URI handler classified every non-binary file as text, which shadowed the executor branch already sitting in the same function. Opening a file:// URL that pointed at an executable shell script dropped it into the editor instead of running it.

    Solution

    Added a pure classify_open_file_action helper that decides the action up front, then routed runnable shell scripts to the executor path the rest of the agent flow already used. Detection gates on the user-execute bit plus a known extension or a shebang on Unix, and on extension alone on Windows, where there is no execute bit.

    Impact

    Merged upstream. 13 unit tests across both layers, +312/-4 over 3 files. Non-executable scripts still open in the editor, so behaviour changes only for files the user actually marked executable.

    Rust
    Systems Programming
    Cross-Platform
    Unit Testing
    Open Source
    View PR #9503
    Upstream Fix in Warp, the Rust Terminal
    Featured Project

    Streamlined 10K+ Monthly Investor Updates

    Boosting VC portfolio tracking efficiency by 72%

    Problem

    VCs and analyst firms struggled to track portfolio company updates efficiently, losing critical insights

    Solution

    Built AI-powered inbox with automated metric parsing, EditorJS templates, and Typst PDF generation

    Impact

    10K+ monthly updates processed, 39+ firms onboarded, 72% efficiency improvement in portfolio tracking

    Amazon SES
    EditorJS
    Typst
    AI Parsing
    NestJS
    Streamlined 10K+ Monthly Investor Updates
    Impact

    Eliminated Manual Cap Table Management for 70+ Companies

    Replacing Excel chaos with automated equity & portfolio systems

    Problem

    Companies managing equity and cap tables through error-prone Excel spreadsheets

    Solution

    Architected end-to-end cap table, equity distribution, and portfolio management platform

    Impact

    70+ companies automated, eliminated manual Excel updates, zero cap table errors

    React
    TypeScript
    PostgreSQL
    Equity Management
    Eliminated Manual Cap Table Management for 70+ Companies
    Cost Savings

    Slashed Infrastructure Costs by 60%

    Boosting system performance while reducing operational expenses

    Problem

    SQL bottlenecks limiting application scalability and driving up infrastructure costs

    Solution

    Architected migration to NoSQL with zero downtime and performance optimization

    Impact

    44% performance improvement, 60% cost reduction, zero downtime migration

    Azure Table Storage
    Node.js
    Migration Scripts
    Performance Optimization
    Slashed Infrastructure Costs by 60%

    Technical Expertise

    Agent frameworks and AI platform architecture on top of a full-stack foundation, from MCP servers and multi-agent orchestration down to the infrastructure they run on.

    AI & Agent Engineering

    Agent Frameworks

    Claude Agent SDK
    Multi-Agent Orchestration
    LangGraph
    Subagent Fleets
    Skills & Plugins

    Protocols & Runtimes

    MCP
    A2A
    AWS Bedrock
    AgentCore
    RAG & Knowledge Bases
    Agent Memory

    Agent Safety & Evals

    Skill Security Scanning
    Prompt Injection Defense
    Agent Identity & OIDC
    Eval Harnesses
    Context Observability

    Frontend Mastery

    Frameworks

    React
    Next.js
    Angular
    Redux

    Styling

    Tailwind CSS
    SCSS
    Material-UI

    Performance

    Web Vitals
    Code Splitting
    Framer Motion

    Backend Excellence

    Languages

    TypeScript
    JavaScript
    C#
    .NET

    Frameworks

    NestJS
    Express.js
    .NET Core

    Databases

    PostgreSQL
    Redis
    Azure Table Storage

    DevOps & Tools

    Cloud

    Azure
    AWS

    Version Control

    Git
    GitHub

    Development

    VS Code
    Docker
    Linux
    Figma

    Achievements & Recognition

    Open Source Contributor

    Merged upstream fix in Warp (Rust)

    3x Employee of the Month

    Augmento Labs

    5⭐ HackerRank

    Problem Solving

    Founding Member

    India's First Equity Secondary Market

    CodeChef President

    SVVV Chapter Founder

    Let's Connect

    Open for technical consulting, speaking engagements, leadership opportunities, and collaborative projects across innovative technology domains.

    Get In Touch

    Location

    Bangalore, India

    Open For:

    • Technical Leadership & Architecture Consulting
    • Speaking Engagements & Technology Workshops
    • Advisory Roles for Technology Startups
    • Collaborative Innovation Projects

    Send a Message

    Download Resume

    Get a detailed overview of my experience, skills, and achievements.