I've spent the past few years as a Machine Learning Scientist and Engineer. At Melonfrost, a synthetic biology startup, I employed deep generative modeling techniques such as Generative Flow Networks (GFlowNets) in a reinforcement/active learning setting and machine learning guided evolutionary algorithms like Adalead (Adapt-with-the-Leader) with the aim of directing the evolution of microbial populations within the company’s high-throughput system of bioreactors. These methods were originally developed to optimize DNA and protein sequence design as well as for applications in drug discovery. In addition to the ML research effort, I designed bespoke biophysical simulation benchmarks, utilized Ray to automate benchmark testing and hyper-parameter optimization on Amazon Web Services, and tracked experiments and model configs with Weights & Biases. As a Ph.D. student in the Department of Theoretical and Computational Chemistry at the University of Cambridge, I developed molecular dynamics simulations to make fundamental theoretical contributions towards understanding non-equilibrium nano-scale flow. Projects included some of the first non-equilibrium calculations of nanoscale flow profiles induced by temperature and chemical potential gradients. Subsequently, as a postdoctoral associate at the Massachusetts Institute of Technology, I used methods from information theory to quantify immune cell signal discrimination between healthy and infected cells. I also developed kinetic Monte Carlo methods to optimize vaccination protocols against highly mutable pathogens. All of this work resulted in numerous publications.

Raman Ganti

Senior Machine Learning Scientist with Ph.D. in Computational Chemistry and Post-Doctoral experience in computational immunology.

Research Expertise

Condensed Matter Physics

Atomic and Molecular Physics, and Optics

Electrical and Electronic Engineering

Biophysics

Surfaces and Interfaces

Electrochemistry

Spectroscopy

About

Legacy Map

Full View

Publications

Molecular Simulation of Thermo-osmotic Slip

Physical Review Letters

2017

Diffraction phase microscopy: monitoring nanoscale dynamics in materials science [Invited]

Applied Optics

2014

How the T cell signaling network processes information to discriminate between self and agonist ligands

Proceedings of the National Academy of Sciences

2020

Measuring the Nonuniform Evaporation Dynamics of Sprayed Sessile Microdroplets with Quantitative Phase Imaging

Langmuir

2015

Pressure gradients fail to predict diffusio-osmosis

Journal of Physics: Condensed Matter

2018

Microscopic Marangoni Flows Cannot Be Predicted on the Basis of Pressure Gradients

Physical Review Letters

2017

Mechanisms underlying vaccination protocols that may optimally elicit broadly neutralizing antibodies against highly mutable pathogens

Unknown Venue

2020

Vibrational and structural signatures of the crossover between dense glassy and sparse gel-like attractive colloidal packings

Physical Review E

2014

Dynamic Clustering Regulates Activity of Mechanosensitive Membrane Channels

Physical Review Letters

2020

Characterizing microdroplet evaporation using diffraction phase microscopy

2014 IEEE Photonics Conference

2014

Hamiltonian Transformation to Compute Thermo-osmotic Forces

Physical Review Letters

2018

Education

University of Cambridge

Ph.D., Chemistry / January, 2018

Cambridge

University of Pennsylvania

MS, Physics / September, 2013

Philadelphia, Pennsylvania, United States of America

University of Pennsylvania

Ph.D., Physics / September, 2013

Philadelphia, Pennsylvania, United States of America

Experience

Melonfrost

Senior Machine Learning Scientist / March, 2022 — August, 2023

Researching and implementing SOTA model-based deep reinforcement/active learning methods such as Adalead, Generative Flow Networks (GFlowNets) with Bayes Opt, and Model Predictive Control for optimal control of selection pressures within company’s bioreactors. Building production-quality (documented, tested, maintainable) python package of machine learning algorithms for strain engineering/optimization. Designing biophysical simulations of evolutionary dynamics and fitness landscapes to simulate bioreactor and test optimization methods. Developing in house simulation benchmarks, utilizing Ray to automate benchmark testing and hyper-parameter optimization on AWS, and logging output and model configs to wandb.

Deep Alpha

Machine Learning Engineer / March, 2021 — March, 2022

Built a cloud-based algorithmic trading system using deep learning methods. Adapting attention-based sequence to sequence models for time series forecasting. Researching deep reinforcement learning ‘actor-critic’ techniques for dynamic portfolio optimization. Finalist at Innospark Ventures AI pitch competition.

Massachusetts Institute of Technology

Post-Doctoral Associate / April, 2018 — April, 2021

Project: Quantifying non-equilibrium behavior using neural network estimators. Keywords: neural networks, machine learning. Project: Designing optimal vaccination protocols. Keywords: kinetic Monte Carlo, information theory, chemical master equations. Project: Understanding how immune cells discriminate between healthy and infected cells. Keywords: ordinary differential equations, systems biology modeling, kinetic proofreading, channel capacity, mutual information.

University of Cambridge

Ph.D. Candidate / January, 2014 — January, 2018

Project: Deriving the theoretical origins of thermo- and diffusio-osmotic flow. Keywords: molecular dynamics, non-equilibrium thermodynamics.

Join Raman on NotedSource!

Join Now

At NotedSource, we believe that professors, post-docs, scientists and other researchers have deep, untapped knowledge and expertise that can be leveraged to drive innovation within companies. NotedSource is committed to bridging the gap between academia and industry by providing a platform for collaboration with industry and networking with other researchers.

For industry, NotedSource identifies the right academic experts in 24 hours to help organizations build and grow. With a platform of thousands of knowledgeable PhDs, scientists, and industry experts, NotedSource makes connecting and collaborating easy.

For academic researchers such as professors, post-docs, and Ph.D.s, NotedSource provides tools to discover and connect to your colleagues with messaging and news feeds, in addition to the opportunity to be paid for your collaboration with vetted partners.