Meghal Donde Pradhan

ProductAIDataUser Experience

I build data and AI products that turn complex information into useful decisions — with evidence, trust, and human judgment built in.

View Projects Read My Work
  • MBA, BU Questrom
  • MCA + BS Computer Science
  • Certified Scrum Product Owner
  • Google AI Professional
  • Anthropic AI Systems
Portrait of Meghal Donde Pradhan

Independent projects

Products and experiments I conceived, designed, and built independently, outside of employer work.

AI Product · Independently Created

Havable: When the answer is no, what then?

A child picks something off the shelf and the label says no. Every allergy app answers that and stops. Havable scans what they picked up and then does the part that actually takes the time: alternatives on this shelf, or a recipe whose ingredients you can buy on the same trip. No model is involved in the allergen decision. Built end to end.

Role
Product · Design · Builder
Context
Built end to end
Status
Live, with a public repository
  • FastAPI
  • Postgres + pgvector
  • Gemini multimodal embeddings
  • Gemini Flash
  • Deterministic allergen gate
  • Open Food Facts (ODbL)
Try it Read the case studyRead the product thinkingView the repository

AI Product · Independently Created

SiftReviews: Is This Review Worth Trusting?

Highlight a review and SiftReviews quotes the phrases that make it read formulaic, and the ones that read first-hand. The site check is verified against Google Safe Browsing, the writing reading is a model estimate, and the two are shown separately so you know which is which. Built end to end.

Role
Product · Design · Builder
Context
Built end to end
Status
Live on Chrome Web Store
  • Chrome Extension (MV3)
  • FastAPI
  • Google Safe Browsing
  • Gemini
  • Selected-text scanning
  • Privacy by design
Install from Chrome Web Store Try the demoRead the case studyView the repository

Writing & Research

Published analysis

Threat-actor research, internet measurement analysis, and essays on working with data. Earlier pieces carry the byline Meghal Donde.

  1. Cyberattacks on France Are Rising—Here’s What the AIDE Data Shows

    Global Cyber AllianceArticle2026

    Published as Meghal Donde

    A threefold rise in attacks against French networks, a Mirai-linked outbound surge from French infrastructure, and where the traffic originates.

  2. Salt Typhoon Across the Internet: What AIDE Honeypots Reveal About a Persistent State-Linked Campaign

    Global Cyber AllianceFull Report2025

    Published as Meghal Donde

    Two years of telecom-decoy telemetry examined against the Salt Typhoon espionage campaign, with mitigation recommendations for network operators.

  3. Dark Pink APT Campaigns Through the Lens of AIDE Telemetry

    Global Cyber AllianceArticle2025

    Published as Meghal Donde

    Attack sessions targeting government, defense, and education organizations across Southeast Asia and APAC, including command-and-control via webhook abuse.

  4. Recurrence Is Not Persistence: Why Repeated Signals Can Mislead Data Teams

    MediumEssay2026

    Published as Meghal Donde Pradhan

    A repeated signal on a dashboard is not evidence that a problem is ongoing. Three questions worth asking about concentration, confounds, and corroboration before an engineering team acts on one.

All writing and research

About

Professional story

I work across product, AI, data, and user experience, turning technical systems and complex datasets into interfaces and workflows people can actually make decisions with. Much of that work has been in cybersecurity and internet measurement.

My background spans software engineering, analytics, data products, partner-facing work, and 0→1 product definition. More recently, I have been focusing on AI product workflows, evaluation, evidence, uncertainty, and human review.

The useful question is rarely what a model can predict. It is what someone is supposed to do next.

Earlier professional publications and public work may appear under the name Meghal Donde.

More about my background

Credentials

Engineering depth, business training, current AI practice

The three things the work above depends on. I came up through computer science rather than into product from the side, and the AI credentials are recent by design.

Business & Strategy

  • MBABoston University Questrom School of Business
  • Beta Gamma SigmaTop 20% of graduating class

Engineering Foundation

  • Master of Computer ApplicationsUniversity of Mumbai
  • B.S., Computer ScienceUniversity of Mumbai

AI & Product

  • Scrum Alliance Certified Scrum Product Owner badgeCertified Scrum Product OwnerScrum Alliance
  • Google AI Professional CertificateGoogle
  • Building AI-Powered SystemsAnthropic
  • Advanced Google AnalyticsGoogle

Professional Experience

Where I have worked

Selected contributions, not a full career history. Employer programs are described only in terms of my contributions.

  1. Oct 2021 – Present

    Senior Software Engineer, Data Management & Analytics

    Global Cyber Alliance

    Co-Chair, Domain Trust Data Working Group

    • Built and evolved large-scale cybersecurity data and analytics systems processing 3M+ daily events across 200+ sensors.
    • Defined measurement, scoring, data, and product requirements across AIDE Internet Pollution Index and Domain Trust work.
    • Reduced AWS OpenSearch cost by 40% and improved query performance by 28% through index, shard, and storage design changes.
    • Worked with research institutions, registries, registrars, and other partners to translate user needs into data access, reporting, and product improvements.

  2. 2017 – 2018

    Co-owner & Lead Software Developer

    Pinac Solutions

    Industrial software, HMI and analytics work.

  3. 2007 – 2012

    Head of Software Development

    On-site Analysis

    Engineering leadership and platform modernization.

Earlier roles across healthcare, industrial automation, edtech, and software engineering.

Consulting & Collaboration

Available for selected engagements

Available for selected engagements involving AI product strategy, data products and analytics, cybersecurity data, and decision workflows.

AI product strategy

Framing what a model should and should not decide, and designing the workflow around it.

Data products & analytics

Measurement frameworks, analytical data models, and reporting people actually use.

Cybersecurity data & measurement

Turning high-volume security telemetry into analysis that supports decisions.

Evaluation & decision workflows

Evidence handling, uncertainty, and where human review belongs.

Discuss a Project