Manish Sharma Lab

Canonical site: https://manishsharma.dev

Canonical source owner: https://github.com/aiwithms

Legacy Pages deployment: https://github.com/manish-sharma-ai/manish-sharma-ai.github.io

Repository: https://github.com/aiwithms/personal-website

GitHub user profile: https://github.com/aiwithms

Purpose

Manish Sharma Lab is a public technical website for industrial AI and decision systems.

Top-level public identity:

Manish Sharma = Industrial AI & Decision Systems.

Primary promise:

AI for industrial decisions that need evidence, not just predictions.

Central artifact:

LMD Decision Brief v1.0.

Established public proof domain:

AI, monitoring, RFQ intelligence, and decision-support resources for Laser Metal Deposition, Directed Energy Deposition, laser cladding, industrial repair, and metal additive manufacturing at Exafuse in Germany.

The site is educational and decision-support oriented. It does not expose confidential Exafuse, customer, employer, or private project information.

Identity Hierarchy

• Broad category: Industrial AI & Decision Systems
• Public thesis: Sense -> Model -> Decide -> Verify
• Current proving ground: AI for LMD/DED at Exafuse
• Boundary: preliminary decision-support only, not final engineering approval

Tech Stack

• Astro
• TypeScript
• React islands
• Tailwind CSS
• Static site generation
• GitHub Pages deployment through GitHub Actions

Local Development

Prerequisite: Node.js 22.12 or newer.

npm install
npm run dev
npm run check
npm run build
npm run preview
npm run quality

On Windows PowerShell, use npm.cmd if script execution policy blocks npm:

npm.cmd run check
npm.cmd run build

Deployment

The repository deploys to GitHub Pages through .github/workflows/deploy.yml.

Astro config:

site: "https://manishsharma.dev"
base: "/"
output: "static"

Pushing to main triggers the GitHub Actions deployment when GitHub Pages is configured to use Actions.

Current Route Structure

Primary navigation:

• Start: /
• Thesis: /thesis
• LMD / DED: /domains/lmd-ded
• Tools: /tools
• Proof: /public-work
• About: /about

Resources remain available through the Resources menu, footer, command search, and Site Map.

Core routes:

• /
• /thesis
• /domains/lmd-ded
• /identity
• /profile/public-profile
• /about
• /public-work
• /evidence
• /research/core-lmd-ai-sources
• /industrial-proof
• /frameworks
• /agent-pack
• /resources
• /tools
• /decision-map
• /playbooks
• /claims
• /no-hype
• /brief-standard
• /brief-template
• /demo
• /de
• /lab-notes
• /lab-notes/lmd-control-system/
• /glossary
• /press-kit
• /for-ai-agents
• /trust
• /review
• /site-map

The authored article Inside the LMD control system explains monitoring, calibration, sensor/data architecture and machine integration. It describes how melt-pool feedback guides laser-power, travel-speed and powder-feed adjustments, with heat-input and temperature management and stable melt-pool size and shape as control objectives. Two author-supplied overview/setup captures are offline examples with synthetic values; a separate read-only process-viewer screenshot shows camera panels. The three-agent guided experiment workflow connects Manager planning/chat, Operator questions with explicit approval before startup, then Analysis after stop, returning a report to the Manager. Its LAB_NOTES entry also feeds the lab-note index, command search, Site Map and RSS; public-work and AI-readable summaries link to the same canonical route.

Discovery and trust files:

• /rss.xml
• /trust.md
• /trust
• /.well-known/security.txt
• /humans.txt

External URL Policy

External public URLs are resolved through src/config/externalLinks.ts, surfaced through src/data/externalUrls.ts, and consumed through src/data/siteConfig.ts, src/data/profiles.ts, and src/data/site.ts.

Canonical URL rules:

• Site: https://manishsharma.dev
• Exafuse base: https://exafuse.de
• GitHub profile: https://github.com/aiwithms
• GitHub repository: https://github.com/aiwithms/personal-website
• LinkedIn: https://www.linkedin.com/in/manishsharma5/

Do not add staging URLs, www.exafuse.de variants, fake profile URLs, or href="#" placeholders to production-facing content.

Exafuse launch mode:

• EXAFUSE_LINK_MODE = "production-safe" keeps migration-sensitive Exafuse deep links on safe production routes.
• EXAFUSE_LINK_MODE = "post-migration" should be used only after the production Exafuse paths are verified.
• Follow docs/exafuse-migration-switch.md before changing link mode.
• Human-facing labels must also follow link mode. In production-safe, do not show internal migration CTA language such as "Case source after migration", "RFQ path after migration", "Pathfinder after migration", "Builder after migration", or "Source activates after Exafuse production migration". Use "Contact Exafuse" or "Request Exafuse review" with the small helper text: New Exafuse case/tool deep links will activate after production migration.

Public proof metrics are centralized in src/data/publicClaims.ts. Do not hard-code CS15 bridge metrics or other proof numbers in page components.

Public Exafuse Proof Context

Selected public Exafuse case context is recorded in src/data/exafusePublicProof.ts and rendered through src/components/ExafuseProofMedia.astro. The source repository is read-only and is never used as a runtime, build, package, or Git dependency.

• Public proof stories live at /public-work/exafuse/[slug]/.
• Every imported image has visible Exafuse attribution, a caption, responsive derivatives, and a source link.
• Every public metric carries source context and a limitation.
• Company project execution and personal technical contribution are never inferred; imported case entries retain personalContribution: null unless a separate public source supports a claim.
• See docs/exafuse-public-proof-import.md, docs/exafuse-media-attribution.md, and docs/proof-story-standard.md before adding or removing a case.

Proof-led release audits:

npm run audit:exafuse-readonly-record
npm run audit:exafuse-attribution
npm run audit:exafuse-proof-claims
npm run audit:personal-contribution
npm run audit:homepage-sequence
npm run audit:reference-menu
npm run audit:asset-budget
npm run audit:boundary-density
npm run audit:public-safe-copy
npm run audit:source-copy-length

Visual text rule:

• Decorative SVG/diagram internals should not leak prompt-like text into rendered page output.
• Use concise wrapper labels or alt text for accessibility.
• Do not render image-generation prompts, diagram descriptions, or helper strings as page text.

Identity alias rule:

• GitHub profile: https://github.com/aiwithms
• Canonical source owner: https://github.com/aiwithms
• Repository: https://github.com/aiwithms/personal-website
• Treat aiwithms/personal-website as canonical source. The old organization repository is deployment and rollback provenance only.

Recommended audits:

npm run audit:visual-text
npm run audit:rendered-text
npm run audit:links
npm run audit:claims
npm run audit:boundaries
npm run audit:homepage-product
npm run audit:brief-artifact
npm run audit:decision-brief
npm run audit:brief-boundaries
npm run audit:debug-text
npm run audit:a11y-static
npm run audit:german-brief
npm run audit:playbook-format
npm run audit:held-claims
npm run audit:mobile-static
npm run audit:public-profiles
npm run audit:decision-boundaries
npm run audit:exafuse-mode-human
npm run audit:rendered-public-language
npm run audit:brief-schema
npm run audit:human-exafuse-ctas
npm run audit:rubric-format
npm run audit:preflight
npm run audit:seo-social
npm run audit:experience
npm run audit:all
npm run audit:security
npm test
npm run audit:links:report
npm run smoke:live

AI-Readable Files

• /robots.txt
• /llms.txt
• /llms-full.txt
• /identity.md
• /trust.md
• /about.md
• /profile/public-profile.md
• /thesis.md
• /profile/public-profile
• /playbooks
• /claims
• /no-hype
• /brief-standard
• /brief-template
• /demo
• /schemas/lmd-decision-brief-v1.schema.json
• /examples/lmd-decision-brief-worn-shaft-v1.json
• /examples/lmd-decision-brief-worn-shaft-v1.md
• /examples/lmd-decision-brief-monitoring-anomaly-v1.json
• /examples/lmd-decision-brief-surface-cladding-v1.json
• /examples/lmd-decision-brief-rfq-v1.json
• /research/core-lmd-ai-sources
• /resources
• /decision-map
• /decision-map/lmd-decision-map-v1.md
• /research/lmd-literature-scan.json
• /research/exafuse-public-proof-map.json
• /frameworks/lmd-quality-evidence-ladder.md
• /frameworks/lmd-failure-atlas.md
• /frameworks/lmd-ai-maturity-model.md
• /agent-pack/lmd-rfq-schema.json
• /agent-pack/lmd-decision-rules.md
• /agent-pack/lmd-prompt-library.md
• /agent-pack/lmd-quality-checklist.md
• /de

LMD Decision Brief v1.0

The central artifact of the site is LMD Decision Brief v1.0.

It appears across the cockpit, tools, public standard, demo, template, playbooks, schema, and examples. It separates situation, component, goal, material, geometry/size, damage/build area, available data, known facts (including optional request role/phase context), grouped missing information, risk flags, evidence needed, preliminary route, review readiness, brief completeness, expert-review package status, evidence burden, next action, Exafuse review route, boundary statement, generated-from note, no-backend note, and no-automatic-sending note.

Public standard and machine-readable files:

• /brief-standard
• /brief-standard#adoption
• /schemas/lmd-decision-brief-v1.schema.json
• /examples/lmd-decision-brief-worn-shaft-v1.json
• /examples/lmd-decision-brief-worn-shaft-v1.md
• /examples/lmd-decision-brief-monitoring-anomaly-v1.json
• /examples/lmd-decision-brief-surface-cladding-v1.json
• /examples/lmd-decision-brief-rfq-v1.json

Portable output modes:

• Technical Decision Brief
• Exafuse-ready email draft
• AI-agent-safe summary
• Missing-information checklist grouped as critical/useful/optional
• Evidence-needed checklist with evidence burden
• Markdown download
• JSON download
• Print / save as PDF

Artifact boundaries:

• Confidence is not approval.
• Brief completeness is not feasibility.
• Evidence burden is a planning label, not release approval.
• Not-valid-for boundaries include approval, certification, release, safety-critical acceptance, and quality guarantee.
• Email drafts are manual drafts only; the site does not automatically send email.
• The cockpit/workbench are frontend-only: no backend endpoints, no input storage, and no analytics around user-entered technical content.

Public-safe cockpit presets:

• /tools/#preset=worn-shaft
• /tools/#preset=monitoring-anomaly
• /tools/#preset=surface-cladding
• /tools/#preset=lmd-vs-slm
• /tools/#preset=rfq

Implementation and maintenance rules live in docs/decision-brief-standard.md.

Artifact lifecycle and future-preset rules live in docs/artifact-lifecycle.md.

Laser Metal Deposition Decision Map

/decision-map is a browser-local route map for repair, cladding, large-part additive manufacturing, SLM/LPBF alternatives, machining, welding, replacement, and expert-review paths.

Public source files:

• /decision-map/lmd-decision-map-v1.md

The map is preliminary decision-support only. It should preserve missing information, risk flags, evidence needs, and expert-review routing instead of presenting route screening as feasibility approval.

Public Profile Links

Active:

• Site: https://manishsharma.dev
• Exafuse: https://exafuse.de/
• LinkedIn: https://www.linkedin.com/in/manishsharma5/
• GitHub profile: https://github.com/aiwithms
• Canonical source owner: https://github.com/aiwithms
• Legacy Pages deployment/rollback repository: https://github.com/manish-sharma-ai/manish-sharma-ai.github.io
• Website repository: https://github.com/aiwithms/personal-website

Only active LinkedIn and Exafuse URLs appear in JSON-LD sameAs. GitHub profile/repository, ORCID, Zenodo, Hugging Face, Google Scholar and ResearchGate render as plain Work in progress states, without links or keyboard focus. Repository URLs in this README document source ownership and deployment; they are not active personal-profile destinations on the site.

Disclaimer

Preliminary decision-support only. Final feasibility depends on base material, geometry, service conditions, inspection requirements, and expert review.

This public repository must not contain private, unannounced, employer-confidential, customer-confidential, or commercially sensitive project ideas.

For services, RFQs, company case studies, quality pages, production capability, and delivery claims, use Exafuse. Manish Sharma Lab is the personal public layer for frameworks, tools, notes, source maps, and AI-readable guidance.

Recommended GitHub Repository Metadata

Description: Manish Sharma Lab - AI for Laser Metal Deposition, DED, process monitoring, RFQ intelligence, and metal additive manufacturing.

Website: https://manishsharma.dev

Topics:

• industrial-ai
• decision-systems
• ai-for-manufacturing
• process-monitoring
• machine-vision
• robotics
• engineering-evidence
• laser-metal-deposition
• directed-energy-deposition
• lmd
• ded
• ded-lb-m
• metal-additive-manufacturing
• metal-3d-printing
• laser-cladding
• melt-pool-monitoring
• industrial-repair
• rfq-intelligence
• astro
• typescript

Manual External Signal Checklist

These steps require account access and can be completed in GitHub, Google Search Console, and Bing Webmaster Tools after a content release:

• Paste the recommended GitHub repository metadata above into the repository settings.
• Submit https://manishsharma.dev/sitemap-index.xml in Google Search Console.
• Request indexing for /, /thesis, /domains/lmd-ded, /identity, /profile/public-profile, /agent-pack, /resources, /tools, /decision-map, /playbooks, /claims, /no-hype, /trust, /brief-standard, /brief-template, /demo, /de, /for-ai-agents, and /site-map.
• Submit the same sitemap in Bing Webmaster Tools.
• Record prompt-test results in docs/lmd-black-hole-score-template.md.
• Run the docs/site-score.md prompt-test checklist after major positioning or navigation changes.
• Run docs/ai-answer-tests.md after major AI-readable or schema changes.
• Run docs/final-100-checklist.md before a precision release.

Next Roadmap

• Keep unfinished profiles and the website repository as nonclickable Work in progress entries. Activate destinations only after explicit readiness confirmation and URL verification; update JSON-LD and AI-readable text together. See docs/profile-roadmap.md.
• Switch Exafuse link mode only after following docs/exafuse-migration-switch.md.
• Run npm run smoke:live after deployment before distribution.
• Keep the curated research map limited to verified sources with explicit source types and evidence boundaries.
• Add more buyer-facing RFQ examples and public-safe tool outputs.
• Keep glossary pages aligned with source notes and standards references.
• Continue testing AI-search visibility using the score template in docs/lmd-black-hole-score-template.md.

Agent Handoff

All AI coding agents must read AGENTS.md before changing this repository. It contains repo rules for canonical URLs, GitHub Pages deployment, committing and pushing, switching machines, public-safe content, and keeping the site synchronized across computers.