Skip to content
View manufaujdar's full-sized avatar

Achievements

Achievement: Pull Sharkx2Achievement: Pair ExtraordinaireAchievement: Quickdraw

Achievements

Achievement: Pull Sharkx2Achievement: Pair ExtraordinaireAchievement: Quickdraw

Block or report manufaujdar

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
manufaujdar/README.md
A constellation connecting healthcare, trustworthy AI, research, and creative systems Portfolio status across trustworthy AI, clinical interfaces, translational research, and creative systems

Manu Faujdar

Healthtech builder · Trustworthy AI researcher · Clinical workflow designer

Master LLM · AI Gateway · Smart Glasses · LinkedIn

◈
TRUST
privacy · safety · evaluation		 ⌁
CLINICAL
workflow · devices · care		 ◌
RESEARCH
evidence · validation · translation		 ✦
CREATE
products · media · interfaces

I build systems that make complex technology more observable, safe, and useful—from AI routing and privacy controls to clinical devices, translational research tools, and creative AI products.

Selected public work

◈ AI Routing Gateway

Explainable model and tool routing with deterministic decisions, specialist-agent workflows, and explicit safety boundaries.

Explore repository →

⌁ Smart Glasses

Vendor-neutral healthcare smart-glasses research with capability contracts, simulators, safe-stop behavior, and workflow experiments.

Explore repository →

More public projects

◈ AI Firewall
Fail-closed sensitive-data inspection for outbound AI calls.

◌ SystemBench
Reproducible evaluation for complete AI systems.

⌁ Reflection AI
Consent-aware personalization and memory framework.

✦ SignalForge
Content intelligence and evaluation dashboard prototype.

✦ VideoViber
Agentic, spec-driven video creation workspace.

◇ Prompt Enhancer
FastAPI service for improving text prompts.

Engineering focus

clinical workflow · intended use · privacy controls · explainable routing · failure modes · validation planning · synthetic testing

I turn ambiguous health-technology problems into bounded systems: define the user and non-use case, map risks, make decisions observable, test failure paths, and document what the evidence does—and does not—support.

◈ Trustworthy AI

  • allow / redact / block controls
  • metadata-only audit events
  • explainable model and tool routing
  • explicit human authorization gates

⌁ Clinical prototypes

  • vendor-neutral capability contracts
  • disconnect and safe-stop behavior
  • privacy boundaries
  • synthetic, offline-testable workflows

◌ Validation

  • evidence mapping
  • intended-use definition
  • confounder and provenance planning
  • repeat-trial evaluation

✦ Delivery

  • needs framing
  • prototype requirements
  • implementation-risk mapping
  • clinical / product / data / engineering coordination

The wider portfolio

AI safety · model evaluation · clinical workflow · smart devices · biomarkers · digital health · creative AI

The public repositories are the visible edge of a larger portfolio of private product labs and research systems. Private work is intentionally represented as a visual category rather than linked before it has a safe public release, complete documentation, and a reproducible demo.

Evidence boundary: clinical projects shown here are research-stage or synthetic/non-clinical unless explicitly stated in the repository.

Working principles

◈ Observable ⌁ Bounded ◌ Resilient ✦ Memorable
Make decisions inspectable Make evidence and non-use explicit Design for graceful failure Give complex systems a human interface

Now

Turning separate experiments into coherent public flagships with reproducible demos, explicit safety boundaries, and evidence-aware engineering.

Popular repositories Loading

  1. Prompt-Enhancer Prompt-Enhancer Public

    Framework to evaluate, enhance, and optimize prompts before LLM API calls

    Python

  2. VideoViber VideoViber Public

    AI-assisted video vibe generation and editing workspace

    TypeScript

  3. ai-routing-gateway ai-routing-gateway Public

    Explainable AI routing gateway that sends each request to a suitable model or specialist workflow.

    Python

  4. systembench systembench Public

    End-to-end AI system evaluation framework for routing, tools, provenance, safety, latency, operability, and task success.

    Python

  5. ai-firewall ai-firewall Public

    Local-first AI firewall that inspects outbound data and allows, redacts, or blocks sensitive information before cloud AI calls.

    Python

  6. smart-glasses smart-glasses Public

    Research platform for simulating clinical smart-glasses workflows, device contracts, capture, voice, and safe-stop states.

    Python