Harsh Desai

Engineer, aiming for a PhD

About Me

About Me

Who am I?

  • life's purpose: inter-planetary, cosmic evolution, guided by hinduism, ai, and galaxial travel. But, to do this, I'm aiming to get a PhD in Space Instrumentation
  • as a soul-seeker; i found my purpose from Srimad Bhagavatam, transcending the physical with Sri Krishna's teachings
  • Studying hindu vedas & puranas, to seek cosmic purpose from my sanatan dharma
  • i seek discomfort, thrive in awkward travels, backpack in manali mostly, love being uncomfortable
  • rode the crypto wave: made and lost fortunes in trading, still love business stiudies, all the shop talk, fundamentals and growth studies. mon NSE valuation setup
  • built a: rocket design optimizer cum flight simulator as saas; college's tax perks kept us student-focused
  • i love a mix of techno and kirtans, prolly a weird combo, but it helps me stay sane.

Check out my Resume.

Education

University of Michigan - Ann Arbor, MI, USA
Master of Engineering (M.Eng) - Space Systems Engineering, CGPA: 3.807/4.0 (Aug'24 - Present)
Coursework: Space Instrumentation, Magnetosphere, Control for Aerospace Vehicles, Space Weather Modeling, Remote Sensing, Space Systems Design and Mgmt., Space Physics

Vellore Institute of Technology, Vellore - Vellore, IND
B.Tech. - Mechanical Engineering (Jul'19 - Jul'23)
Focus: Computational Engineering & Simulations - CFD, CAE
Special Achievers Award Recipient #

Achievements

Special Achievers Award recipient 2019-22 #
• Recognized as a Special Achiever by VIT for outstanding international representation and performance.

SA Cup 2021 (organized by ESRA) #
• The team secured 23rd rank globally, being 5th in Asia-Pacific (out of 75 teams) at our very first attempt.

CANSAT 2022 (Tethered Payload) - Organized by NASA & AAS #
• Secured a personal best worldwide rank of 7th (out of 42 teams).
Projects

Projects

Z-score-based Medium Frequency Pairs Trading Setup #

Trading System capable of reading the correlation between Nifty and BankNifty Implied Volatalities. Implementing XGBoost to predict the profit centers based on the time series data of their volatilities. Based on Time Series Multivariate Trading strategy. Performing Statistical arbitrage with Nifty and BankNifty options.

Agricultural Predictive Insurance Modelling #

Focus has been on building predictive insurance models for agriculture. This involves deploying an ML model for comprehensive weather prediction using time-series data, to enhance weather forecasting to refine insurance rate predictions for the agricultural sector. Additionally, I am leveraging GeoSpatial data with QGIS and image processing to analyze crop and soil health for insurance risk assessment. The goal is to optimize premium adjustments based on real-time climate insights and the health of crops and soil.

Flight Simulator #

Flight Simulator Enhancement: We enhanced a flight simulator to address a 15% apogee discrepancy found in standard simulations. By integrating Barrowman equations with Computational Fluid Dynamics (CFD) analysis via PyFluent, our system now accurately predicts rocket trajectories and stability, overcoming the limitations of conventional models like OpenRocket. This allows for optimized rocket design considering real-world perturbations.

Ti/Te Investigation from MHD & hybrid models on CCMC (Feb'25 – Apr'25)

Paper: Link

Multi-fluid sim: Utilizing SWMFv23 with MAGE on CCMC, figuring out limitations in resolving independent ion and electron energy equations within a global multi-fluid MHD framework—to backtrace Ti/Te spatial distributions across bow shock and magnetosheath domains.

Had to further investigate using HYPERS- hybrid model on CCMC. Downstream reconnection‐driven E∥J∥ heating reveals peak electron temperature excursions preceding ion maxima, with ΔTi - Te oscillations dominated by convective transport, evidencing cross‐species energy partitioning during shock events.

FAA Class II Solid Motor Sounding Rocket (@ SA Cup) (Aug'20 – Jul'23)

Project Technical Report: Link

Built a Sounding Rocket targeting 10,000 ft. using a solid rocket motor capable of reaching 0.9 Mach. Successfully launched it this year at Spaceport America, USA. Dealt with exterior ballistics Design + Simulations, while coordinating mechanical system designs with Avionics, mitigating design conflicts, and harmonizing delivery timelines. Developed a PID-controlled payload with real-time pitch correction and integrated advanced avionics using dual-deployment altimetry and GPS tracking. Built an SRAD FSW + Flight Simulator capable of predicting trajectory based on design optimisation done via back-end CFD Simulation.

CANSAT 2022 #

Built a Can-sized satellite shall consist of a container, which will descend a tethered payload to 10 meters in 20 seconds, while being attached to the container by a 10-meter-long tether. The CanSat will be launched to an altitude ranging from 670 meters to 725 meters above the launch site and deployed near apogee to start descending using a parachute at a rate of 15 m/s, with the payload maintaining the orientation of a video camera pointing in the south direction, 45 degrees downward to assure terrain is in the video. secured a personal best worldwide rank of 7th!

CANSAT GUI #

Designed to interact with an onboard flight computer in real-time. Its main objectives included receiving and displaying telemetry data, sending commands for tasks such as telemetry control and sensor calibration, and integrating various functionalities through Python's PyQt5 framework. The GUI utilized tools like Matplotlib for graphing, OpenStreetView Map with pyqtlet for mapping, and libraries such as csv, time, pyserial, paho-mqtt, and datetime for file generation, timing, serial communications, MQTT protocol integration, and UTC time setting, respectively.

CANSAT 2021 (Aug'20 – Jul'21)

Critical Design Review: Link

Simulated an atmospheric re-entry vehicle in the CANSAT mission with two Maple Seed-inspired Monowing payloads for live telemetry; developed custom SRAD Ground Control Software. - Created and optimized a Mono-wing design using transient simulation for flow analysis, employing a MATLAB script for maximum torque and efficiency.

CANSAT 2020 #

mission was to design a delta wing glider capable of descending down from the container in a helical motion with a radius of 250 meters collecting sensor data for one minute and remain above 100 meters after being released. After which a parachute will deploy to bring down the glider at a rate of 10 m/s. successfully completed the design of a delta wing glider with my team in the span of 1 month.

Work Experience

Work Experience

Rivera Event-Directional Discontinuity Correlational Study (May'25 – Present)

Advisor: Mojtaba Akhavan-Tafti | University of Michigan

Understanding how Directional Discontinuities evolve in the Rivera conjunction. To understand if the discontinuities' temporal matching and field variable feature matching reveal correlation to the heating of the solar wind stream. Developing on the TS method to understand analytically ΔB change. To understand TS-multi-s/c correlation and also MVA differences in results and categorization.

Enhanced sampling on Solid State Detectors using XADCs on FPGAs (Jan'25 – Jul'25)

Advisor: Prof. Stefano Livi, Solar and Heliospheric Research Group (SHRG) | University of Michigan

Optimizing hard ADC IPs on FPGA fabric, trying "FPGA as ADCs" as well for Silicon Strip Detectors(SSDs). Replacing flash-ADC or Analog processing chains using shaping amplifiers to even higher sampling rates, with finer acquisition control, filtering and pile-up rejection.

Working on Xilinx Zynq architectures, DMA capabilities to improve digitization in XADC and power efficiency of the board. Implementing complete eFPGA optimization Systems Engineering process on MATLAB HDL coder, running Co-Simulation on AMD Vivado - no need for iterative VHDL/Verilog scripting.

Comms Lead - Uranian Orbiter and Probe (Sep'24 – Apr'25)

L3 Harris, Spacecraft Technology (SPACE 582 & 583) | University of Michigan

Leading the Communications Subsystem of the Uranian Orbiter and Probe mission. Currently, focusing on Science Mission design and modeling. Extrapolating the current Uranian MHD models to Science Instrumentation selection and design-for Fields & Particles instrument package optimization.

Developing cost-effective alternative to NASA's UOP Flagship Mission through trade studies across probe deployment, orbital optimization, telemetry allocation, and Ka-band/optical hybrid solutions for 4.5-year science operations.

Project Intern (Jun'24 – Aug'24)

PentaShield Technologies Pvt. Ltd. | Vadodara, IND

Engineering a program capable of running multi-objective optimization of ballistic missile airframe designs, integrating a backend CFD solver, in line with surrogate modeling, and MODOs using only IGES geometry as input.

Automated parametric design workflow using PyFluent to train the model to predict the best possible design output, enhancing efficiency in geometry pre-processing and validation for ANSYS simulations, while integrating PyOptiSLang for optimization.

Predictive Insurance Modeling for Agri-Insurance sector (Nov'23 – Dec'23)

Research Assistant @ Christ University | Pune, IND

Deploying a TensorFlow model for comprehensive weather prediction using time-series data. Objective: Enhance weather forecasting to refine insurance rate predictions for the Agricultural Sector. Leveraging GeoSpatial data with QGIS + image processing to analyze crop & soil for insurance risk assessment. Optimize premium adjustments based on real-time climate insights and crop/soil health.

Projects

Skills

Languages

MATLAB, Python, C++, Verilog(Learning), VHDL(learning)

Space Science + Engineering

AMD Vivado, ANSYS STK, ModelCenter(learning), ANSYS ODTK(learning), SPENVIS, SWMF & Space-weather Modeling(MAGE, SWMF, HYPERS), MATLAB HDL Coder & DSP Toolbox

Data Tools

Tableau, Google Data Studio, QGIS, BigQuery, Google AppsScript, AppsFlyer

Mech Tools

ANSYS(GUI/TUI): Fluent, Mechanical; PyANSYS, OptiSlang, OpenFOAM, SolidWorks, Fusion 360, AutoCAD, Autodesk Inventor, MIDO, Xflr5, RocketPy, OpenRocket

Additional

Git, JavaScript, Node.js, CSS, ReactJS, JAVA, Python(Flask), TensorFlow, PyTorch, LangChain, R, MySQL, PowerBI

Other stuff I did

Contact me
harshdes@umich.edu
+1 7345481080