OPEN TO RESEARCH, ML, AND DEFENSE-TECH ROLES

ML ENGINEER, RESEARCHER & PROBLEM SOLVER

I build machine learning systems, research-driven simulations, and data products that translate complex technical work into measurable real-world impact.

Personal Photo
MACHINE LEARNING • DEEP REINFORCEMENT LEARNING • FLOOD PREDICTION • UAV AUTONOMY • RESEARCH • DATA VISUALIZATION • PYTHON • PYTORCH • MACHINE LEARNING • DEEP REINFORCEMENT LEARNING • FLOOD PREDICTION • UAV AUTONOMY • RESEARCH • DATA VISUALIZATION • PYTHON • PYTORCH •

TECHNICAL SKILLS

Python
React.js
JavaScript
HTML
CSS
Bootstrap
SQL
AWS
Django
Flask
Docker
NumPy
SciPy
scikit-learn
TensorFlow
Keras
PyTorch
Pandas
OpenCV
API Testing

EXPERIENCE

Work Experience

Office of the Relief Commissioner, Govt. of Uttar Pradesh, India

Disaster Mitigation Associate

Uttar Pradesh, India

Sept 2023 - Jan 2024
  • Developed a machine learning-based flood prediction model to support disaster management decision-making.
  • Designed and implemented a visualization dashboard for real-time flood prediction insights.
  • Prepared a comprehensive project report covering methodology, model performance, and recommendations.
  • Created documentation for the ML model, gamification system, dashboard, and data integration pipelines.
  • Generated strategic recommendations to strengthen disaster mitigation planning.
Internship Experience

Indian Space Research Organization (ISRO), Ahmedabad, Gujarat

Research Intern

Ahmedabad, Gujarat

Oct 2022 - Mar 2023
  • Improved image spatial resolution using advanced 2D CNN techniques by merging high- and low-resolution data.
  • Curated 90+ sampling points and synthesized datasets using Synthetic Data Vault and Gretel.ai.
  • Generated and analyzed soil nutrient data with a focus on organic carbon and nitrogen.
  • Contributed to data enrichment workflows for improved soil analysis.

PROJECTS

Featured Project

GNSS-Denied UAV Path Planning using Deep Reinforcement Learning (PPO vs SAC)

A defense-oriented autonomous navigation project focused on enabling UAV path planning without GPS or pre-mapped environments.

100% PPO Success RateMulti-Seed ValidationA* Comparable Performance
Deep Reinforcement LearningPPOSACGymnasiumUAV Autonomy
  • Designed a GNSS-denial simulation framework for autonomous UAV navigation without GPS using deep reinforcement learning.
  • Implemented and compared PPO and SAC in a custom Gymnasium environment.
  • Built a defense-oriented evaluation framework using Success Rate, Average Episode Reward, and Steps-to-Goal.
  • Achieved 100% success rate with PPO across multiple obstacle densities, significantly outperforming SAC.
  • Demonstrated that PPO can match A* path planning performance without requiring GPS or environment maps.
  • Conducted ablation studies and multi-seed validation to ensure robustness and reproducibility.
VIEW PROJECT DETAILS →
Featured Project

Real-Time Flood Prediction Dashboard

A strategic decision-support system developed for the Govt. of UP to predict and visualize flood risks using machine learning.

88% Peak AccuracyReal-Time IntegrationGovt. Deployed
Machine LearningDashboardOptimizationPython
  • Built a machine learning-based flood prediction model to support high-stakes disaster mitigation.
  • Designed an interactive dashboard for visualizing sensor data and prediction insights.
  • Optimized data pipelines to handle real-time hydrological inputs and surface runoff modeling.
VIEW PROJECT DETAILS →
Formal Learning

CERTIFICATIONS

Certificate links can be added later, so I've updated the section content-first and kept the layout ready for future credentials.

Highlights1 Featured Credential

Space systems, communications, and emerging technologies aligned with research-driven work.

01Aug 2023 - Sept 2023

Glue Technologies for Space Systems Ph.D. Summer School

Universita di Trento

Attended an intensive program on space communications and emerging technologies, covering Space 2.0, quantum communications, and satellite networking with global experts and researchers.

Certificate

HACKERRANK CERTIFICATIONS

Validated technical skills and algorithmic proficiency.

HackerRank

Problem Solving (Basic)

HackerRank2023
AlgorithmsData Structures
HackerRank

Python (Basic)

HackerRank2023
PythonControl FlowStrings
HackerRank

SQL (Basic)

HackerRank2023
SQLRelational Databases
HackerRank

Java (Basic)

HackerRank2023
JavaOOPBasics
HackerRank

JavaScript (Basic)

HackerRank2023
JavaScriptES6DOM
HackerRank

CSS (Basic)

HackerRank2023
CSSFlexboxGrid

RESEARCH & PUBLICATIONS

ICAAV 2026Under Review

A GNSS-Denial Simulation Framework for Evaluating Deep Reinforcement Learning in Autonomous UAV Path Planning: PPO vs SAC

  • Proposed a GNSS-denied UAV simulation framework for autonomous navigation without GPS using deep reinforcement learning.
  • Conducted a comparative analysis of PPO and SAC, achieving up to 100% success rate with PPO across varying obstacle densities.
  • Introduced a defense-oriented evaluation framework using Success Rate, Average Reward, and Steps-to-Goal.
  • Demonstrated that PPO matches A* path planning performance without requiring GPS or environmental maps.
IEEEManuscript in Preparation

Quantifying the Downstream Decision Cost of Miscalibrated XGBoost Predictions Under Varying Imbalance Ratios and Dataset Sizes

  • Investigates the impact of probability miscalibration in XGBoost on real-world decision-making under imbalanced data.
  • Introduces Decision Cost Ratio (DCR) to quantify cost increase caused by threshold misalignment.
  • Shows that miscalibration can increase operational cost by up to 18-41% in low-data, high-imbalance scenarios.
  • Evaluates Platt Scaling and Isotonic Regression for improving calibration and downstream decision performance.

ACHIEVEMENTS

Startup Flight 2025

2025

Certificate of Participation in the international startup program organized by NSSC, VietJet Air, Amity Innovation Incubator, and Vietnam Innovation Hub.

Research Presentation at ISRO

Feb 2023 - Mar 2023

Presented research on hyperparameter tuning of various models, including methodology, flowchart, findings, and final conclusions.

Internship Offer - DRDO

Dec 2022

Secured an internship opportunity from DRDO, reflecting strong fit for defense and applied research environments.

Conference Participation - AAV 2023

Dec 2022

Attended Autonomous Air Vehicles - Technologies and Applications (AAV 2023), jointly organized by ADE-DRDO and Design Division-AeSI.

HOW I WORK

I enjoy combining rigorous experimentation, strong documentation, and practical product thinking. My work style is analytical, collaborative, and impact-focused.

  • • RESEARCH-FIRST EXECUTION
  • • ROBUST MODELS WITH CLEAR EVALUATION
  • • DOCUMENTATION THAT MAKES SYSTEMS USABLE

BUILDING SYSTEMS THAT HOLD UP UNDER PRESSURE

From disaster mitigation to GNSS-denied UAV navigation, I like solving high-stakes problems where reliability, explainability, and performance all matter.

LET'S BUILD SOMETHING MEANINGFUL

I'm open to research collaborations, internships, and full-time opportunities in ML, AI, data, and autonomous systems.

EMAIL
LINKEDIN
GITHUB