Roshith Mittakolu

Roshith Mittakolu

PhD Candidate, Mechanical Engineering
Thermal-Fluid Systems · Scientific Computing
Experimental Testing and Validation
University of Maryland

Designing, Modeling, and Testing
Next-Generation Cooling Systems
for Data Centers and Electronics.
From governing equations and numerical methods to simulation software, system-level design, and hands-on experimental testing and validation.

I am a mechanical engineering PhD candidate at the University of Maryland. My work focuses on thermal management of data centers and electronics, sitting at the intersection of thermal-fluid systems, numerical methods, scientific computing, and experimental testing and validation.

My PhD revolves around three major projects, two ARPA-E COOLERCHIPS efforts and a Northrop Grumman collaboration. In MOSTCOOL, I develop system-level simulation tools for two-phase data center cooling. Through a collaboration with the University of Florida, I extend these tools to study a novel membrane-assisted heat sink concept. The Northrop Grumman project adds an experimental component for designing and characterizing a novel cold plate with encapsulated phase change materials.

I see myself as a mechanical/thermal engineer. Software became a core part of how I work because the problems I study required tools that need to be developed as the existing ones were insufficient. Outside of my core research, as a hobby, I explore how AI and LLMs might eventually change how thermal engineers simulate, reason, and design.

Open to industry roles in data center thermal management, electronics and power electronics cooling, and advanced thermal systems.

Download CVMore about me →
Technical Interests
Data Center Thermal ManagementTwo-Phase CoolingLiquid CoolingElectronics CoolingComputational Heat TransferNumerical MethodsSystem-Level Thermal DesignScientific SoftwarePhysics-Based SimulationExperimental Testing and ValidationElectronics PackagingAI/ML for Thermal Engineering

Education

Where I studied and what I built along the way

University of Maryland, College Park

University of Maryland, College Park

Doctor of Philosophy, Mechanical Engineering

August 2022 – Present · GPA 3.83 / 4.0

Doctoral research in thermal management of data centers and electronics, with emphasis on two-phase cooling, system-level thermal design, scientific simulation tools, and experimental thermal engineering.

Indian Institute of Technology (IIT) Gandhinagar

Indian Institute of Technology (IIT) Gandhinagar

Master of Technology, Mechanical Engineering

July 2019 – June 2021 · CPI 8.6 / 10

Graduate research focused on computational fluid dynamics and numerical methods for compressible flow.

National Institute of Technology (NIT) Calicut

National Institute of Technology (NIT) Calicut

Bachelor of Technology, Mechanical Engineering

July 2014 – May 2018 · CGPA 7.46 / 10

Undergraduate training in thermodynamics, fluid mechanics, heat transfer, and solid mechanics.

Selected Research

Current projects and collaborations

ARPA-E COOLERCHIPS · System-Level Simulation
MOSTCOOL

System-level design and simulation tools development for next-generation data center cooling, with a strong emphasis on two-phase flow network modeling. My branch-first adaptive tearing solver achieves a 25x improvement in convergence robustness over current state of the art.

Two-phase flow networksBranch-first adaptive tearingChip to rack to facilitySole developer
Python
GitHub
NumPy
SciPy
Ansys Fluent
Numba
Scikit-Learn
Linux
VS Code
View project →
ARPA-E COOLERCHIPS · University of Florida Collaboration
Membrane-Assisted Cold Plate Cooling

System-level implications of a hydrophobic membrane-based two-phase cooling architecture for extreme power density racks. Work includes manifold sizing, condenser sizing, pump requirements, and scaling analysis toward 1 MW rack-level systems with the two-phase module I developed as part of MOSTCOOL.

1 MW rack concept studiesMembrane heat sinkSystem design specificationsUF collaboration
Python
VS Code
Linux
SciPy
NumPy
MATLAB
View project →
Northrop Grumman · Modeling + Experiments
Encapsulated Phase Change Slurry Cooling

Combining modeling and ongoing experimental work to study encapsulated phase change material slurry cooling with dielectric fluids for electronics. Defined two novel design-oriented thermal performance metrics and developed a three-regime operating framework for EPCM slurry cooling in dielectric microchannels. Currently designing and characterizing a novel cold plate that enhances the utilization of EPCM particles.

Experimental — ongoingEPCM slurryData acquisitionCode testing
Ansys Fluent
Python
MATLAB
SolidWorks
LabVIEW
Cold Plate Testing
Data Acquisition (DAQ)
View project →
View all research →

Selected Publications

Featured research output

Graphical abstract: Augmenting the heat capacity of dielectric liquids using encapsulated phase change materials
Applied Thermal Engineering · 2026
Augmenting the heat capacity of dielectric liquids using encapsulated phase change materials
R. Mittakolu, V. V. Manepalli, Y. Kim, A. Ateş, C. Pullins, M. T. Barako, S. Graham, P. McCluskey, D. Agonafer
Under review

Defines design-oriented effectiveness metrics and a three-regime operating framework for EPCM slurry cooling in dielectric microchannels.

Ansys Icepak
Python
MATLAB
LaTeX
Illustrator
C++
Graphical abstract: Towards 1 MW data center racks: System-level modeling of a membrane-assisted two-phase cooling architecture
ITherm — Intersociety Conference on Thermal and Thermomechanical Phenomena · 2026
Towards 1 MW data center racks: System-level modeling of a membrane-assisted two-phase cooling architecture
R. Mittakolu, I. Gandikota, D. Agonafer, P. McCluskey, S. Moghaddam
Accepted

System-level modeling of a novel membrane-assisted cold plate concept for extreme power density racks targeting 1 MW scale.

Python
LaTeX
Illustrator
Graphical abstract: Higher-order implicit shock-capturing scheme based on linearization of implicit fluxes for the Euler equations
International Journal of Numerical Methods for Heat & Fluid Flow · 2023
Higher-order implicit shock-capturing scheme based on linearization of implicit fluxes for the Euler equations
R. Mittakolu, S. L. Rani, D. S. Sundaram
Published

Develops a higher-order implicit shock-capturing scheme for the Euler equations using linearization of implicit flux Jacobians.

MATLAB
LaTeX
PowerPoint
JEP
Journal of Electronic Packaging · 2026
Physics of failure approach for reliability analysis of liquid cooled data centers
I. Gandikota, R. Mittakolu, V. Raavi, P. McCluskey
Under review

Physics-of-failure methodology for reliability analysis of liquid cooling systems in data centers.

Python
View all publications →

Technical Toolkit & Skills

Methods, platforms, and tools I use in research and engineering

Finite Element Method
Finite Volume Method
Finite Difference Method
Nonlinear Solvers
Sparse Linear Solvers
Time Integration
Python
C++
MATLAB
NumPy
SciPy
Pandas
Numba
Scikit-Learn
Anaconda
Jupyter
Spyder
Git
GitHub
Docker
Linux
VS Code
Ansys Fluent
Ansys Icepak
PyAnsys
SolidWorks
ReliaSoft
Data Acquisition (DAQ)
DMA Testing
DSC Calorimetry
Cold Plate Testing
Uncertainty Analysis
LabVIEW
Matplotlib
Origin Pro
Illustrator
LaTeX
PowerPoint
Canva
Jira
Obsidian
Zotero
LLM Agents
RAG Pipelines
OpenClaw
LlamaIndex
Physics-Informed ML
Workflow Automation
Finite Element Method
Finite Volume Method
Finite Difference Method
Nonlinear Solvers
Sparse Linear Solvers
Time Integration
Python
C++
MATLAB
NumPy
SciPy
Pandas
Numba
Scikit-Learn
Anaconda
Jupyter
Spyder
Git
GitHub
Docker
Linux
VS Code
Ansys Fluent
Ansys Icepak
PyAnsys
SolidWorks
ReliaSoft
Data Acquisition (DAQ)
DMA Testing
DSC Calorimetry
Cold Plate Testing
Uncertainty Analysis
LabVIEW
Matplotlib
Origin Pro
Illustrator
LaTeX
PowerPoint
Canva
Jira
Obsidian
Zotero
LLM Agents
RAG Pipelines
OpenClaw
LlamaIndex
Physics-Informed ML
Workflow Automation
View full skills breakdown →