data.json

{
  "meta": {
    "title": "My World Map",
    "subtitle": "An interactive, zoomable world of who I am, what I do, who's in it, and what I'm building. Drag to pan, pinch or scroll to zoom, tap a planet to fly in.",
    "center_label": "You",
    "center_sub": "Phanindra Malladi",
    "generated": "2026-08-13",
    "artist": "Your Hermes CEO",
    "source_note": "Derived from L1 memory + second-brain wiki + session history. Tap any planet to fly in and explore its moons."
  },
  "planets": [
    {
      "id": "identity",
      "name": "Identity",
      "color": "#f472b6",
      "size": 1.06,
      "orbit_speed": 0.35,
      "orbit": 1.0,
      "summary": "Who I am at the core — the center of gravity of everything here.",
      "intro": "This is the identity planet: the person everything else orbits around. Knowing this is what lets every project, every decision, and every relationship make sense.",
      "moons": [
        {
          "label": "Phanindra Malladi",
          "detail": "AI Team Lead, transitioning to Product Owner. Hyderabad, India.",
          "body": "The operating system of this whole map. Everything you see — the ventures, the work, the people, the memories — all orbit this single identity. I am first an engineer and architect of autonomous systems, and second a builder of products and companies.",
          "points": [
            "Name: Phanindra Malladi",
            "Location: Hyderabad, India",
            "Path: Voice support → SaaS founder → Enterprise AI lead"
          ]
        },
        {
          "label": "Core skill: Enterprise AI architect",
          "detail": "Builds autonomous 'Digital FTEs' — full-code agent systems that do real operational work inside regulated enterprises.",
          "body": "The thread that ties my career together: taking heavy, repetitive, high-stakes operational work and replacing it with agent systems built in code, not demos. The arc has been from running my own SaaS for a decade to leading the design of a whole factory for such agents.",
          "points": [
            "Autonomous agents, not one-off automations",
            "Regulated industries: Life Sciences, Security, HR",
            "Full-code agents — written, tested, verified, shipped"
          ]
        },
        {
          "label": "Approach",
          "detail": "Autonomy-first: decide and execute routine steps on my own; pause only at safety gates.",
          "body": "How this whole system runs. I built my orchestration to be maximally autonomous — routine pipeline steps execute without asking. I only stop to ask at real gates: publishing something live, promoting it, spending money, or anything touching legal or reputation.",
          "points": [
            "Routine steps: auto-execute",
            "Gates: publish / promote / spend / legal",
            "The human stays the CEO; the agent runs the factory"
          ]
        },
        {
          "label": "The goal",
          "detail": "Turn a ~Rs 3,000/month VPS into a multi-agent side-hustle revenue engine.",
          "body": "The north star for all of this. A single modest VPS running agent orchestration 24/7, generating real revenue from the AIdentify Gumroad store with as little daily human involvement as possible.",
          "points": [
            "Infra: ~Rs 3k/month VPS",
            "Revenue: AIdentify Gumroad store",
            "Mode: fully autonomous unless a gate says pause"
          ]
        }
      ]
    },
    {
      "id": "work",
      "name": "Work",
      "color": "#60a5fa",
      "size": 1.16,
      "orbit_speed": 0.27,
      "orbit": 1.0,
      "summary": "IntelligenzIT — leading the Enterprise Digital Worker Factory.",
      "intro": "My day job and where the autonomous-agent ideas I explore at night get hardened into production reality.",
      "moons": [
        {
          "label": "IntelligenzIT — AI Team Lead",
          "detail": "Apr 2026–present. Lead of the Enterprise Digital Worker Factory.",
          "body": "I lead the design and delivery of a 'Digital Worker Factory' at IntelligenzIT (Hyderabad). It's a productized architecture for shipping repeatable autonomous agents rather than one-off automations — a factory line for AI workers that run real operations inside large, regulated client enterprises.",
          "points": [
            "Started: April 2026",
            "Role: AI Team Lead → moving to Product Owner",
            "Domain: Enterprise Digital Worker Factory"
          ]
        },
        {
          "label": "Productized architecture",
          "detail": "core/ → vertical → client template → agent. Repeatable, not bespoke.",
          "body": "The factory's core idea: instead of building each agent from scratch per client, we layer a shared core, vertical-industry knowledge, per-client template configuration, and then the agent itself. Each layer is reusable, so shipping the next agent gets faster and more predictable each time.",
          "points": [
            "core/ (shared foundation)",
            "→ vertical (industry knowledge)",
            "→ client template (configuration)",
            "→ agent (the shipped worker)"
          ]
        },
        {
          "label": "Chandra (Digital Cloud Engineer)",
          "detail": "An agent I own covering 5 KRAs: cost, security, compliance, performance, reliability. Success bar ≥80% recall on seeded misconfigurations.",
          "body": "Chandra is my flagship agent at IntelligenzIT — a Digital Cloud Engineer that watches cloud environments and catches real problems. It's measured hard: on a test corpus of deliberately seeded misconfigurations, it must recall at least 80%. That's an evaluation bar, not a promise — a real, verifiable success criterion.",
          "points": [
            "5 KRAs: cost, security, compliance, performance, reliability",
            "Success bar: ≥80% recall on seeded misconfigs",
            "Delivered end-to-end + demoed live to an enterprise Life Sciences client"
          ]
        },
        {
          "label": "DPI-LS scoring framework",
          "detail": "A 0–100 framework to rate Digital Workers across 7 dimensions. 276 passing tests.",
          "body": "DPI-LS is how we grade an agent's worth in a regulated setting. It scores a Digital Worker across 7 dimensions — productivity, quality, execution, governance, risk, validation, and cost — and reduces it to a single 0–100 number. Backed by 276 passing tests, it turns 'is the agent good?' into a measurable, defensible score.",
          "points": [
            "7 dimensions: productivity, quality, execution, governance, risk, validation, cost",
            "Score range: 0–100",
            "276 passing tests"
          ]
        },
        {
          "label": "Pod leadership",
          "detail": "I run a pod of engineers + interns: hiring, KRA definition, code review, real-data verification.",
          "body": "Beyond the code, I own people and quality. That means hiring engineers and interns, defining their KRAs, reviewing their code, and — critically — verifying claims against real data rather than taking 'it works' at face value.",
          "points": [
            "Hiring + KRA definition",
            "Code review",
            "Real-data verification of claimed completions"
          ]
        },
        {
          "label": "Client-facing",
          "detail": "Solution scoping, live demos, and post-demo architecture pivots.",
          "body": "The factory sells to real clients. That part of my job is scoping what they need, demoing the agents live, and then pivoting the architecture based on how the demo actually lands. It keeps the engineering honest.",
          "points": [
            "Solution scoping",
            "Live demos",
            "Post-demo architecture pivots"
          ]
        },
        {
          "label": "Domains",
          "detail": "Regulated industries: Life Sciences (GxP), IT Security, HR.",
          "body": "The factory's verticals. Each is heavily regulated, which is precisely why autonomous agents are valuable there — and why evaluation, governance, and validation are non-negotiable. Never name real clients publicly; use archetypes like 'a top-20 pharma'.",
          "points": [
            "Life Sciences (GxP-aware)",
            "IT Security",
            "HR"
          ]
        }
      ]
    },
    {
      "id": "stack",
      "name": "Stack",
      "color": "#34d399",
      "size": 0.98,
      "orbit_speed": 0.4,
      "orbit": 1.0,
      "summary": "Tools and tech I build autonomous agents with.",
      "intro": "My technical arsenal — the concrete libraries, clouds, and patterns behind every agent in this map.",
      "moons": [
        {
          "label": "LangGraph",
          "detail": "StateGraphs for multi-agent orchestration and stateful agent workflows.",
          "body": "My orchestration foundation. LangGraph gives me explicit graphs of nodes and state transitions — the right tool when an agent isn't a single prompt but a pipeline of steps that carry state, branch, and coordinate multiple agents.",
          "points": [
            "StateGraphs",
            "Multi-agent orchestration",
            "Branching + persistent state"
          ]
        },
        {
          "label": "AWS",
          "detail": "Bedrock, RDS, S3, IAM, Terraform — the cloud backbone.",
          "body": "Everything runs on AWS. Bedrock for LLM inference, RDS for data, S3 for storage, IAM for fine-grained permissions, and Terraform to manage it all as infrastructure-as-code. Also supports air-gapped / local-LLM deployments for clients who can't send data to the cloud.",
          "points": [
            "Bedrock, RDS, S3, IAM, Terraform",
            "Air-gapped + local-LLM capable",
            "Infrastructure as code"
          ]
        },
        {
          "label": "Provider-abstracted model layer",
          "detail": "A single build_chat_model() factory that switches Bedrock / local LLM.",
          "body": "A small but vital pattern: never hard-code a model provider. One factory function builds the chat model, so the same agent code runs against Bedrock in the cloud or a local LLM on-prem — whichever the client's compliance posture allows.",
          "points": [
            "One build_chat_model() factory",
            "Switch between Bedrock and local LLM",
            "Client-compliant model selection"
          ]
        },
        {
          "label": "Data & UI",
          "detail": "Postgres for storage, Streamlit for quick internal UIs.",
          "body": "Postgres holds the state and data behind the agents; Streamlit gives fast, lightweight interfaces for demos and internal tooling without a heavy frontend build.",
          "points": [
            "Postgres: durable storage",
            "Streamlit: quick UIs"
          ]
        },
        {
          "label": "Claude Code CLI",
          "detail": "Agentic coding help for building and shipping.",
          "body": "I use agentic coding tools (Claude Code CLI) to move faster on the actual agent codebases.",
          "points": [
            "Agentic coding assistant",
            "Faster shipping"
          ]
        },
        {
          "label": "Agent evaluation",
          "detail": "Recall harnesses and scoring engines — how I prove an agent works.",
          "body": "The discipline that makes all this credible: I don't say an agent works, I measure it. Recall harnesses test whether it catches the problems it should (Chandra's ≥80% bar), and scoring engines like DPI-LS turn quality into a number.",
          "points": [
            "Recall harnesses",
            "Scoring engines",
            "Regulated-environment validation"
          ]
        }
      ]
    },
    {
      "id": "ventures",
      "name": "Ventures",
      "color": "#fbbf24",
      "size": 1.12,
      "orbit_speed": 0.22,
      "orbit": 1.0,
      "summary": "Companies and side projects — the founder arc and current builds.",
      "intro": "From a decade as a SaaS founder to today's autonomous-agent side projects. This planet traces that arc.",
      "moons": [
        {
          "label": "AscendSaaS — Founder (2016–2026)",
          "detail": "SaaS growth/marketing product scaled to 500+ paying users. Wound down to go all-in on enterprise AI.",
          "body": "AscendSaaS was a SaaS growth and marketing product I founded and ran for a full decade. It grew to 500+ paying users — a real, proven business. In 2026 I wound it down deliberately to go all-in on enterprise AI. That decade is my founder/operator credibility.",
          "points": [
            "Founder, 10-year run (May 2016–May 2026)",
            "500+ paying users",
            "Wound down for enterprise AI"
          ]
        },
        {
          "label": "AIdentify — Founder (2024–)",
          "detail": "The autonomous digital product factory → Gumroad store. My primary side project and the point of this whole system.",
          "body": "AIdentify is where my auto-autonomy happens: an autonomous factory that researches ideas, scores them, writes PRDs, builds products, QAs them, pushes to GitHub, drafts Gumroad listings, and prepares promo — all by agents. This 'My World Map' itself is an AIdentify-style build. Note: publicly it's framed as a personal project, not a business or second revenue stream.",
          "points": [
            "Pipeline: research → score → PRD → build → QA → GitHub → Gumroad → promo",
            "Driven by the @aidenify_ceo agent",
            "Primary Gumroad side project",
            "Framed publicly as a side project"
          ]
        },
        {
          "label": "OpsForge",
          "detail": "AI agent skill marketplace — BYOK model, verified skill packs, managed hosting. Solves 'agents are hard to extend safely.'",
          "body": "OpsForge is the marketplace idea: verified, safe skill packs plus managed hosting, using the customer's own keys (BYOK). It addresses the real pain that agents are powerful but dangerous to extend. Built: a monorepo, a CLI, an installer, a LangGraph ReAct core, three starter skills, and a Docker/FastAPI provisioner.",
          "points": [
            "BYOK (bring your own key) model",
            "Verified skill packs + managed hosting",
            "Built: monorepo, CLI, installer, ReAct core, 3 starter skills, provisioner"
          ]
        },
        {
          "label": "AgentsFactory / Hermes",
          "detail": "Personal LangGraph orchestrator + autonomous opportunity-scanning pipeline. Internal use only.",
          "body": "Before (and alongside) AIdentify, I built a personal LangGraph orchestrator with an autonomous pipeline that scans for opportunities. It's the seed of the automation I later productized.",
          "points": [
            "Personal orchestrator",
            "Autonomous opportunity scanning",
            "Internal use"
          ]
        },
        {
          "label": "Knoah Solutions — Voice Support (2012–2015)",
          "detail": "My start: a Voice Support Officer. The origin of the journey.",
          "body": "The earliest chapter. I started in voice support at Knoah Solutions from 2012–2015 — the ground floor of the arc from that role, through founding a SaaS, to leading enterprise AI today.",
          "points": [
            "Role: Voice Support Officer",
            "2012–2015",
            "Start of the journey"
          ]
        }
      ]
    },
    {
      "id": "aidentify",
      "name": "AIdentify",
      "color": "#a78bfa",
      "size": 1.12,
      "orbit_speed": 0.3,
      "orbit": 1.0,
      "summary": "The autonomous product factory — my main active build.",
      "intro": "The system this whole operation runs on. AIdentify is the autonomous factory that takes an idea to a shipped, gated product with minimal human input.",
      "moons": [
        {
          "label": "The pipeline",
          "detail": "research → score → PRD → build → QA → GitHub verify → Gumroad draft → promo drafts → kaizen.",
          "body": "The assembly line. Agents research ideas, score them against criteria, write a PRD, build the product, QA it, verify the GitHub push actually landed, draft the Gumroad listing, prepare promo, and then run kaizen (continuous improvement) on the cycle. Each step hands off to the next automatically.",
          "points": [
            "research",
            "score",
            "PRD",
            "build",
            "QA",
            "GitHub verify",
            "Gumroad draft",
            "promo drafts",
            "kaizen"
          ]
        },
        {
          "label": "CEO bot: @aidenify_ceo",
          "detail": "The Hermes agent running the factory 24/7 on the VPS.",
          "body": "This very assistant. @aidenify_ceo is a Hermes agent running 24/7 on a VPS, acting as the AIdentify CEO orchestrator. It loads skills on demand, stores prefs in L1 memory, grows the second-brain wiki, and drives the factory forward autonomously.",
          "points": [
            "Runs 24/7 on VPS",
            "Drives every pipeline stage",
            "Loads skills on demand"
          ]
        },
        {
          "label": "Integrations",
          "detail": "Gumroad · GitHub (aidendify) · GHL CRM + Social Planner · Ocoya for X.",
          "body": "The factory's toolbelt. Gumroad for the store, GitHub for shipped code under the aidendify org, GoHighLevel for CRM and social planning, and Ocoya for publishing to X. Each integration has a skill defining exactly how to use it safely.",
          "points": [
            "Gumroad: product store",
            "GitHub: aidendify org",
            "GHL: CRM + Social Planner",
            "Ocoya: X publishing"
          ]
        },
        {
          "label": "Approval gates",
          "detail": "Pause only for publish / promote / spend / legal. Everything else runs.",
          "body": "The safety rails. The factory is autonomous, but it never crosses four lines without a human: publishing something live, promoting it, spending money, or anything with legal/reputation weight. Everything short of those executes on its own.",
          "points": [
            "publish",
            "promote",
            "spend",
            "legal"
          ]
        },
        {
          "label": "Social OS",
          "detail": "Brand @xAIdentify + personal @malladiphani — dual-account X presence.",
          "body": "AIdentify talks to the world through two X accounts: the brand (@xAIdentify) and my personal handle (@malladiphani). Managed via the social lane files and scheduled crons (brand posts, personal pulse).",
          "points": [
            "Brand: @xAIdentify",
            "Personal: @malladiphani",
            "Scheduled via social queue + crons"
          ]
        },
        {
          "label": "Root path & artifacts",
          "detail": "/opt/data/aidentify/ — config, products, PRDs, reports, skills, kanban.",
          "body": "Where it all lives on the VPS. A structured folder that keeps the factory's truth: config, shipped products, PRDs, reports, reusable skills, and a kanban of state.",
          "points": [
            "config.yaml",
            "products/",
            "prds/",
            "reports/",
            "skills/",
            "kanban"
          ]
        }
      ]
    },
    {
      "id": "people",
      "name": "People",
      "color": "#f87171",
      "size": 0.92,
      "orbit_speed": 0.42,
      "orbit": 1.0,
      "summary": "Who's in my orbit — team, clients, and the humans in the loop.",
      "intro": "The people planet: the humans who orbit with me, whether I lead them, sell to them, or answer to them.",
      "moons": [
        {
          "label": "Engineers + interns (the pod)",
          "detail": "The team I lead — I hire, define KRAs, review code, and verify claims against real data.",
          "body": "My direct team. Working with them means more than writing code: it's hiring, setting clear KRAs, reviewing pull requests, and — most importantly — checking that what they say is done is actually done, against real data.",
          "points": [
            "Hiring",
            "KRA definition",
            "Code review",
            "Real-data verification"
          ]
        },
        {
          "label": "Enterprise clients",
          "detail": "Life Sciences, mobility, agri-industrial — always abstracted in public.",
          "body": "The clients the factory serves. In any public-facing material they are anonymized to archetypes — 'a top-20 pharma', 'a European mobility OEM', 'an Indian agri-industrial conglomerate' — never real names, per the safety rule.",
          "points": [
            "Never name real clients",
            "Use archetypes instead",
            "Regulated-industry buyers"
          ]
        },
        {
          "label": "PVR",
          "detail": "The building-in-public approver — gates legal/reputation publishing decisions.",
          "body": "A key human in the loop. Before anything sensitive ships publicly (like the full portfolio), PVR gives the building-in-public sign-off. This is one of the few people whose approval I'm wired to stop and wait for.",
          "points": [
            "Gate: building-in-public",
            "Authority: publish/legal decisions"
          ]
        }
      ]
    },
    {
      "id": "personal",
      "name": "Personal Life",
      "color": "#2dd4bf",
      "size": 0.9,
      "orbit_speed": 0.44,
      "orbit": 1.0,
      "summary": "Life beyond the code — the public footprint and daily reality.",
      "intro": "The personal planet: where the professional world meets day-to-day life and the public internet.",
      "moons": [
        {
          "label": "Building in public",
          "detail": "AIdentify + portfolio live at phanindramalladi.online (Hostinger → GitHub Pages).",
          "body": "The public face of my work. My portfolio lives at phanindramalladi.online, hosted on GitHub Pages with the domain at Hostinger. It's framed around me as an AI engineer and team lead — never as 'Founder of a business' — with AIdentify shown as a personal side project.",
          "points": [
            "URL: phanindramalladi.online",
            "Hosting: Hostinger domain → GitHub Pages",
            "Positioned as AI engineer/team lead, AIdentify as side project"
          ]
        },
        {
          "label": "Personal channels",
          "detail": "Primary working life in Hyderabad; Slack AIDENTIFY + Telegram as my comms.",
          "body": "The infrastructure of my communication with this system. Home is a Slack workspace (AIDENTIFY), with Telegram as a secondary backup, and this whole operation runs 24/7 from the VPS.",
          "points": [
            "Based: Hyderabad",
            "Primary: Slack AIDENTIFY",
            "Backup: Telegram"
          ]
        }
      ]
    },
    {
      "id": "memory",
      "name": "Memory",
      "color": "#e879f9",
      "size": 1.0,
      "orbit_speed": 0.32,
      "orbit": 1.0,
      "summary": "How my world is stored — the system that powers this very map.",
      "intro": "The memory planet is meta: it's the storage system that generated this whole map, and the reason it can stay live and growing.",
      "moons": [
        {
          "label": "L1 Hot memory",
          "detail": "MEMORY.md + USER.md — injected every turn: autonomy gates, home channels, paths, secrets locations.",
          "body": "The always-on, tiny memory budget that's loaded into every conversation so I never forget the essentials: how much autonomy you want, where your home channels and paths are, and where secrets live. It's small on purpose — that's why it's the hot layer.",
          "points": [
            "Injected every turn",
            "Holds gates, channels, paths",
            "Deliberately tiny"
          ]
        },
        {
          "label": "L2 Second-brain wiki",
          "detail": "/opt/data/wiki — entities, concepts, diagrams, log (an Obsidian vault).",
          "body": "The compounding knowledge layer. Beyond the tiny hot memory, everything durable lives here as linked pages — entities, concepts, diagrams (including this map), and an append-only log. It's Obsidian-compatible, so it can be browsed like a real vault.",
          "points": [
            "Path: /opt/data/wiki",
            "Entities, concepts, diagrams, log",
            "Obsidian-compatible"
          ]
        },
        {
          "label": "L3 Skills",
          "detail": "/opt/data/skills — procedural 'how to act' knowledge, loaded on demand.",
          "body": "Skills are my procedural memory: step-by-step how-to for recurring tasks, loaded only when relevant. This keeps the always-on memory tiny while still having deep, reusable procedures for anything I do more than once.",
          "points": [
            "Loaded on demand",
            "Step-by-step procedures",
            "Keep hot memory small"
          ]
        },
        {
          "label": "L4 Artifacts",
          "detail": "/opt/data/aidentify + GitHub + Gumroad — shipped-work truth.",
          "body": "The ground truth of what's actually done. Shipped products, code on GitHub, live Gumroad listings. L4 is where 'did we ship it?' is answered — and where I verify a push really landed before claiming success.",
          "points": [
            "Path: /opt/data/aidentify",
            "GitHub org: aidendify",
            "Gumroad listings"
          ]
        },
        {
          "label": "Live update",
          "detail": "This map regenerates when memory changes — you're looking at the living copy.",
          "body": "The point of all of this: because memory feeds a data.json that this page polls, the map is alive. When I learn a new fact about you, it can be folded in and the page updates. The solar system you see is a snapshot of the living system.",
          "points": [
            "data.json → page polls every 30s",
            "Memory changes → map updates",
            "You're looking at the living copy"
          ]
        }
      ]
    }
  ]
}