About Me

Hello, I'm Megh! I try to keep this page up-to date, but I sometimes fall behind. You can contact me at:

• Email: meghss at proton dot me
• Linkedin: meghsuthar

Work History

Principal, Boltanic ~ Present

Boltanic is my consultancy company, which I use for applied AI consultancy work. I'm not a typical consultant. For better or worse, I like being in the trenches.

Sr. Software Engineer, FOSSA ~ 2021 - 2024

FOSSA provides an open-source management platform (for vulnerability and license compliance) for companies like Ericsson, Uber, Slack, Ford, etc. I architected and led the development of key projects – container scanning, dependency analyzers and early reachability analysis.

• Architected and spearheaded the initial native container scanning product, enabling the scanning of Linux (Ubuntu, CentOS, etc.) container images for CVEs and compliance issues related to installed packages and dependencies.

• Implemented (for Java) and authored (for Python) the initial code flow analysis to identify if a vulnerability associated with a dependency is reachable in the customer's code. This was a cool functionality, essentially creating a call graph (caller, callee) relationship via parsing source code and imposing the dependency's call graph to create a finalized call graph and infer if the function associated with the vulnerability is reachable.

• Participated in sales and customer calls, and led/spoke on webinars as a technical expert on relevant topics.

• Developed and maintained a static analyzer for dependency analysis for the following package managers:

  • JS Ecosystem (npm, pnpm, yarn)
  • Python Ecosystem (poetry, pdm)
  • Java Ecosystem (gradle, some of maven)
  • .Net
  • PHP
  • Yocto/Bitbake
  • Dart
  • Swift
  • R
  • Fortran
  • Nim

• Collaborated with the web app team to implement associated UI changes and contributed to various quality-of-life functionalities across the product. Setup new services, and endpoints.

• Initiated several DevEx processes – telemetry (from CLI), automated API docs, and sample projects for automated testing, which have evolved and are actively used.

Head of Machine Learning, Absorb ~ 2019 - 2021

Absorb offers learning management solutions to enterprises with over 30 million daily active users. I was brought on to grow a machine-learning arm from scratch, close competitive gaps, and deliver a differentiated product.

I also worked as a manager and team lead prior to the head of ML role.

• Led a team of engineers, analysts, and QAs for all machine learning and search-related programs and product features across Absorb’s product.

• Architected and developed award-winning products: Intelligent Assist, Automated Transcriptions, Report Suggestions, Auto-tagging, Content-generation and Content Recommendation (closing competitive gaps and delivered product differentiation). Refer to: https://www.absorblms.com/features/lms-ai/

• Contributed to technical committees: backend-practice group, architecture group, and established Python development practices. Also fostered a culture of experimentation (A/B), data-driven decision-making, and operational excellence.

CTO, Emagin (acquired by Innovyze) ~ 2016 - 2019

Emagin provides a real-time monitoring and operational intelligence platform for industrial utilities/plants with an AI platform. Emagin was venture-backed by Tencent and has active deployments across the globe.

• Led the managerial team and teams of software engineers, engineers, product, and support, totaling fourteen (14) staff members.

• Set the strategic vision and technological direction for B2B product offerings, developed an ambitious roadmap, and implemented it to win million-dollar contracts in local and European markets.

• Owned critical relationships with technology partners and vendors. Acted as a technical spokesperson for whitepapers and sales materials, and presented at conferences, sales meetings, and to investors.

• Developed core IP (as part of the initial engineering team):

  • Developed model-serving architecture for hundreds of ML models.
  • Developed a data-ingestion pipeline to process and perform near-real-time optimization at scale.
  • Architected similarity search, alarming, and virtual sensor systems from time-series data at scale.
  • Productionized optimization scheduling algorithms for aeration scheduling and pump optimizations.
  • Developed the initial user interface and control dashboard in React.
  • Managed key PoC projects at early stages – worked with regulatory, engineering, and IT stakeholders.

System Analyst, Environment Canada ~ 2013

• Developed software for nutrient loading analysis in C# for conservation authority and internal teams.

• Data mined U.S. soil datasets to train classifiers (using C4.5) for interpolation of Canadian soil profiles.

IT Analyst, Purolator ~ 2013

• Developed dashboards in QlikView for IT and business teams, focusing on pallet calculations and analysis.

• Automated initial employee auditing (which previously took quarters!) with VBA and Excel.

• Helped build an employee security database by automating the assignment of photographs to employee profiles with Python. Purolator used to take pictures of employees/contractors and store them in a computer at the associated terminal. These sometimes had partial names or initials. Used SVM for classification.

Frontend Engineer, eProf Education ~ 2012

• Developed a web app in Rails (with Ruby) and jQuery, allowing students and instructors to join and create webinars.

Publications, Talks, etc.

• Containers and Open Source License Compliance for Practicing Law Institute
• Patent: Predictive Modelling and Control for Water Resource Infrastructure (pending)
• IDF Curve Lookup for the Province of Ontario
• Extreme Precipitation Time Trends in Ontario, 1960–2010
• OWWA seminar (2017)

Education

• AWS DevOps Engineer Certificate
• MASc, University of Waterloo (dropped out to focus on Emagin)
• BASc, University of Waterloo