Machine Learning Researcher With Professor Bogdan Raita at Georgetown University·2025 – now

Research ML that learns from irregular data

With Professor Bogdan Raita I co-develop Ehrenpreis–Palamodov Gaussian Process models.

Talk to me about
Co-Founder at Georgetown University Business & Healthcare Association·2024 – now

Started McDonough's first healthcare group

I co-founded the Georgetown University Business & Healthcare Association and lead its startup accelerator.

Talk to me about
Georgetown University Alumni & Student Federal Credit Union

Help run credit builder programs for students

Our credit builder programs serve 3,500+ student and alumni members, and I oversee loan disbursements.

Talk to me about
DC Crash Network Recovery Prototype·2026

Recovered DC crash locations, 90.4% accurate

About 20.6% of recent DC crash records had shaky intersection data, which biased hotspot analysis.

Talk to me about
agentmail-context-optimization·2026

Halved email retrieval context, same answers

Retrieval ranks what looks relevant. I tested which email threads the agent needed: only 2 of 4.

Talk to me about
dynamic-mfs-laplace

Wrote two neural PDE solvers in PyTorch

One is a physics-informed network for Poisson's equation; the other learns its own source positions for Laplace's.

Talk to me about
Consultant at Kearney

Helped launch a Saudi wellness initiative

In Dubai I built Excel models testing IRR, NPV and payback across hospitality and residential assets.

Talk to me about
Krishav Gulati

Krishav Gulati⁠

About

I do machine learning research and build data pipelines.

Experiences

  1. Machine Learning Researcher With Professor Bogdan Raita at Georgetown UniversityResearch ML that learns from irregular data

    With Professor Bogdan Raita I co-develop Ehrenpreis–Palamodov Gaussian Process models.

  2. Co-Founder at Georgetown University Business & Healthcare AssociationStarted McDonough's first healthcare group

    I co-founded the Georgetown University Business & Healthcare Association and lead its startup accelerator.

  3. Georgetown University Alumni & Student Federal Credit UnionHelp run credit builder programs for students

    Our credit builder programs serve 3,500+ student and alumni members, and I oversee loan disbursements.

  4. DC Crash Network Recovery PrototypeRecovered DC crash locations, 90.4% accurate

    About 20.6% of recent DC crash records had shaky intersection data, which biased hotspot analysis.

  5. agentmail-context-optimizationHalved email retrieval context, same answers

    Retrieval ranks what looks relevant. I tested which email threads the agent needed: only 2 of 4.

  6. dynamic-mfs-laplaceWrote two neural PDE solvers in PyTorch

    One is a physics-informed network for Poisson's equation; the other learns its own source positions for Laplace's.

  7. Consultant at KearneyHelped launch a Saudi wellness initiative

    In Dubai I built Excel models testing IRR, NPV and payback across hospitality and residential assets.

Follow Krishav for more...

InstagramLinkedInWebsite

© 2026 Delphi · Terms · Privacy

Krishav Gulati

Krishav Gulati⁠

About

I do machine learning research and build data pipelines.

Experiences

  1. Machine Learning Researcher With Professor Bogdan Raita at Georgetown UniversityResearch ML that learns from irregular data

    With Professor Bogdan Raita I co-develop Ehrenpreis–Palamodov Gaussian Process models.

  2. Co-Founder at Georgetown University Business & Healthcare AssociationStarted McDonough's first healthcare group

    I co-founded the Georgetown University Business & Healthcare Association and lead its startup accelerator.

  3. Georgetown University Alumni & Student Federal Credit UnionHelp run credit builder programs for students

    Our credit builder programs serve 3,500+ student and alumni members, and I oversee loan disbursements.

  4. DC Crash Network Recovery PrototypeRecovered DC crash locations, 90.4% accurate

    About 20.6% of recent DC crash records had shaky intersection data, which biased hotspot analysis.

  5. agentmail-context-optimizationHalved email retrieval context, same answers

    Retrieval ranks what looks relevant. I tested which email threads the agent needed: only 2 of 4.

  6. dynamic-mfs-laplaceWrote two neural PDE solvers in PyTorch

    One is a physics-informed network for Poisson's equation; the other learns its own source positions for Laplace's.

  7. Consultant at KearneyHelped launch a Saudi wellness initiative

    In Dubai I built Excel models testing IRR, NPV and payback across hospitality and residential assets.

Follow Krishav for more...

InstagramLinkedInWebsite

© 2026 Delphi · Terms · Privacy

By using this service, you agree to the Terms of Service and Privacy Policy.