Data science · Aerospace engineering

Pramishan Regmi

I build machine learning systems, and I started as an aerospace engineer.

MS in Data Science and BS in Mechanical and Aerospace Engineering. I like problems where physics and data meet, such as predictive maintenance and performance optimization.

Based inNepal
DegreesMS Data Science, BS MAE
Works withPython, SQL, AWS, ANSYS
Open toData science and aerospace roles

Experience

Data Scientist

Achievers Technologies LLC · Sheridan, WY
Dec 2024 – Aug 2025

Data Science Intern

Code Gravity LLC · Dallas, TX
Nov 2023 – Aug 2024

Manufacturing Engineer

Zimmer Biomet · Warsaw, IN
Sep 2022 – Mar 2023

Engineering projects

Senior capstone · MAE 4188 · Team PLATH

3D-printed competition aircraft

Our team designed, printed and flew a fixed-wing aircraft for the 6th Annual 3D Printed Aircraft Competition. The plane gets at most 8 seconds of powered launch, then has to stay airborne inside a 300 ft by 160 ft by 30 ft field. Every airframe part is 3D printed. Only the propeller and electronics are off the shelf.

We started from a baseline glider, built the geometry in OpenVSP, then modeled it in SolidWorks. The wing uses a NACA 63-212 airfoil with an internal I-beam and rib structure sized by treating the wing as a cantilever beam. The tail uses NACA 0012. The airframe is printed in sections with ABS on FDM printers, and the electronics mounts are printed on an SLA printer.

We built three prototypes. The second had weight, center of gravity and print quality problems, because parts came from several printers. For the third we cut about 80 g from the tail, removed a fuselage section to shorten the boom, redesigned the servo mounts and moved to a larger propeller. Flight tests still showed a lateral center of gravity offset and too little thrust, so we ran balance and wind tunnel tests on the next iteration.

34 inwingspan
< 2 lbflying weight
3prototypes built and flown

Tools and methods OpenVSP · SolidWorks · FDM and SLA printing · Structural sizing · Flight testing

Final aircraft
Wingspan34 in
Fuselage length24 in
Wing airfoilNACA 63-212
Tail airfoilNACA 0012
ControlsRudder, elevators
PowerBrushless motor, 2S 300 mAh Li-Po
Propeller12 × 6E, up from 7 × 4E
Static thrust (calculated)about 0.9 lb to 7.7 lb
Print sections6 wing, 4 fuselage
Wall thickness2 bead (FDM), 1 mm (SLA)

MAE 3303 · Supersonic aerodynamics

Supersonic airfoil design

Designed a sharp-edged, two-panel airfoil for Mach 2.5 that maximizes lift-to-drag under thickness constraints. I wrote MATLAB code that solves the oblique shock and Prandtl-Meyer relations to get lift, wave drag and leading-edge pitching moment, then searched angle of attack and panel geometry for the best feasible design.

Feasible designs had an expansion fan on the upper leading edge and a shock on the lower one. Small geometry changes moved the results a lot.

4.26best lift-to-drag
Mach 2.5design condition
6°angle of attack at best

Tools and methods MATLAB · Shock-expansion theory · Optimization

View MATLAB script (.m)

Top feasible designs, chord = 1
AOAxuxltutlcl/cd
6°0.900.600.030.074.26
7°0.970.600.030.074.07
6°0.400.600.040.083.83
6°0.400.600.030.093.62
6°0.400.600.040.103.19

MAE 4342 · Mechanical Design II · Team of 5

Prototype gearbox for an overhead crane

Sized a three-stage spur gear reduction for a 7.5 ton crane. I worked through gear ratios, tooth counts, diametral pitch and face width from bending stress, then designed the shaft and the key. The key is the intentional weak link, so an overload fails the key before it damages the gears.

200:1total reduction
15 hpinput at 1800 rpm
9 rpmoutput speed

Tools and methods Spur gears · AGMA bending and surface stress · Shaft fatigue · Key design

Final gear set
PdTeethFace, inPitch dia, in
Pinion 113140.8751.1
Gear 213700.8755.4
Pinion 28141.3751.75
Gear 38701.3758.75
Pinion 38143.8751.75
Gear 481123.87514

MAE 4344 · Finite element analysis

Hand pump stress analysis

Modeled a hand pump well in SolidWorks (seven parts, 32.5 lb assembly) and analyzed it in ANSYS with structural steel in place of gray cast iron. I compared a normal 200 lbf pumping load with a 20,000 lbf overload for deformation, natural frequency, von Mises stress and factor of safety.

14.8safety factor at 200 lbf
2.56safety factor at 20,000 lbf
590 psipeak stress at 200 lbf

Tools and methods SolidWorks · ANSYS · Modal analysis

ANSYS total deformation plot of the hand pump model with modal frequency chart
ANSYS total deformation and modal frequencies.

Manufacturing processes · Team of 3

Scissor jack for high-volume production

Designed a scissor jack and planned how to make it at 2.5 million units per year. Nearly every part comes from the same 0.25 in galvanized steel sheet, which is sheared, drilled, ground and folded. The trunnions and rod cap are green sand cast. The report has CAD models, engineering drawings, a traveler for every part and a cost breakdown per assembly.

Tools and methods SolidWorks · Process planning · Cost estimation · Design for manufacturing

Rendered scissor jack assembly
Final assembly render.

More coursework

MAE 3314 · Heat transfer

Microprocessor chip cooling

Finite difference model of 2D steady conduction and convection through a chip, heat spreader and motherboard, written as nodal energy balances. I tested whether a higher convection coefficient or an aluminum alloy spreader keeps the chip under its temperature limit. The nodal equations are solved as a matrix system in Excel.

Download Excel workbook

MAE 3314 · Heat transfer

Gas-to-water heat exchanger

Sized a thin copper tube that heats water with hot gas flowing over it. I used Nusselt correlations, iterated film properties and an energy balance to find the tube length, then checked it against the design limit. The Excel workbook sweeps gas temperature (250, 375 and 500 °C), gas velocity and tube diameter (20, 30 and 40 mm).

Download Excel workbook

Skills

Data science and ML

Python, SQL, PySpark, supervised and unsupervised learning, regression, classification, time series forecasting

AI and NLP

TensorFlow, PyTorch, Keras, BERT, spaCy, BERTopic, LLMs, prompt engineering

Data engineering and cloud

Spark, Databricks, ETL pipelines, AWS (S3, Lambda, SageMaker), Azure

Visualization and tools

Tableau, Power BI, MLflow, Airflow, Docker, FastAPI, Git and GitHub

Engineering and simulation

ANSYS, MATLAB, SolidWorks, heat transfer, aerodynamics, mechanical system design

Education

MS Data ScienceUniversity of the Cumberlands, KY · Aug 2024
BS Mechanical and Aerospace EngineeringUniversity of Texas at Arlington · May 2022

Contact

I am open to data science roles and to aerospace work where data and engineering overlap. Email is the fastest way to reach me.

pramishanregmi@gmail.com