Hi, I'm

Jenil Padshala

AI/ML Engineer & Systems Builder

Computer Science engineer focused on machine learning, edge AI, and high-performance systems. I enjoy taking ideas from research papers to real, deployed systems — from autograd engines in C++ to assistive vision devices running on the edge.

Download ResumeGitHubLinkedInEmail

Career

Experience

AIML Research Intern

June 2024 – July 2024

Carnegie Mellon University

Pittsburgh, PA, USA

  • Conducted advanced time-series analysis using LSTM RNNs to analyze ECG data, integrating temporal data with non-temporal data to improve clinical decision-making.
  • Attended Multi-Modal Machine Learning classes and successfully completed graduate-level assignments.
  • Fostered a habit of reading and analyzing research papers to stay current with the latest advancements in AI & ML.

Selected Work

Projects

Scalar Autograd Engine

June 2026
C++PythonPybind11Git
  • Developed a decoupled, lightweight automatic differentiation engine in C++ utilizing custom Value containers linked via weak pointers to prevent cyclic memory references.
  • Implemented a reverse-mode automatic differentiation backend using an iterative Depth-First Search (DFS) topological sort to securely compute gradients over complex multi-variable Directed Acyclic Graphs (DAGs).
  • Supported core arithmetic operations via operator overloading, validating correctness through a robust multi-variable calculus test suite.
  • Exposed the high-performance C++ backend to Python using Pybind11 to enable seamless, high-level interface experimentation and validation against popular deep learning frameworks.

Real-Time Visual Assistance System

Feb 2025 – May 2025
PythonRPi 5YOLOv11nOpenCVVLMs
  • Engineered an edge-AI assistive device on a Raspberry Pi 5 optimized with a Hailo-8 AI Accelerator to provide real-time spatial awareness and multi-modal feedback for visually impaired individuals.
  • Optimized and deployed a concurrent vision pipeline executing object detection (YOLOv11n), multi-object tracking (HailoTracker/JDE), and monocular depth estimation (scdepthv3) delivering 29 FPS at 720p resolution.
  • Developed an offline voice interface using the Vosk toolkit for local wake-word and command speech-to-text transcription, integrated with a Mistral LLM to parse and categorize user intent.
  • Constructed a multi-stage text processing pipeline leveraging OpenCV for image deskewing, EasyOCR for raw character extraction, and LLM API post-processing to clean text anomalies before gTTS generation.
  • Implemented a cloud-connected scene understanding module utilizing a LLaVA-13B vision-language model hosted via Ollama on Azure Cloud to deliver rapid, concise textual environment descriptions.

Time-Series Analysis of ECG

June 2024 – July 2024
PythonLSTMData Pre-processingTime-Series
  • Conducted exploratory data analysis and pre-processed ECG data, addressing missing values and ensuring extraction of high-quality ECG recordings.
  • Developed an LSTM model that achieved 89.1% accuracy in predicting diagnostic super-classes from 12-lead ECG signals.
  • Proposed future steps to integrate non-temporal data with the predictions for enhanced performance.

Toolbox

Technical Skills

Languages

C/C++PythonSQLJavaShell Scripting

ML & Data Science

PyTorchOpenCVSklearnNumPyPandasMatplotlib

Web & Databases

FastAPIStreamlitExpress.jsNode.jsJavaScriptMySQL

Cloud & DevOps

DockerAWSGoogle Cloud Platform

Background

Education

SRM University AP

B.Tech in Computer Science and Engineering · Amaravati, AP

CGPA: 9.06 / 10.0

2021 – 2025

Atmiya Vidya Mandir

AISSCE (12th Grade) · Surat, GJ

95.4%

2020 – 2021

Atmiya Vidya Mandir

AISSE (10th Grade) · Surat, GJ

94.0%

2018 – 2019

Say Hello

Get in Touch

I'm always open to discussing AI/ML, systems engineering, research, or new opportunities. Feel free to reach out.

Emailjenilpadshala@gmail.comGitHubJenilPadshalaLinkedInjenil-padshala
Download Resume