Merishna Singh Suwal

Professional Experience

Collins Aerospace (RTX)

Data Scientist Sr. Analyst & Technical Lead

July 2023 – Present

Leading data science initiatives and technical development for aerospace maintenance solutions.

Lead PI planning sessions and sprint ceremonies, driving agile delivery across multiple data science teams
Orchestrate cross-functional collaboration between engineering, product, and SME teams for NLP-based maintenance solutions
Built custom semantic similarity model pipelines using NER and SBERT with comprehensive MLFlow tracking
Developed internal Plotly dashboard with knowledge graph analysis, enabling data-driven aircraft component insights

Impact: Reduced operational downtime by 80% through automated workflows

University of New Haven

Graduate Research Assistant | Project Lead & ML Engineer

Sep 2022 – Dec 2023

Led research initiatives and managed large-scale data science projects for academic applications.

Authored research paper on "Reinforcement Learning for Epidemic Control" presented at IISE Annual Conference 2023
Led 15-member data science team for ML-based graduate admissions system
Conducted literature reviews and contributed to ML model development from inception to publication

Impact: Achieved 75-80% accuracy in student admissions prediction

Doorstead Inc.

Data Scientist

May 2021 – Mar 2022

Applied machine learning techniques to optimize real estate pricing strategies and market analysis.

Applied ML techniques (CatBoost, Random Forest, Time-Series) for home rental price predictions
Developed Streamlit dashboard for model performance visualization and outlier detection
Leveraged FbProphet for time-series forecasting and market trend analysis

Impact: Achieved 5% sales boost through targeted marketing optimization

Kharpann Enterprises Pvt. Ltd.

CTO & Data Scientist

Feb 2019 – May 2021

Co-founded technology company and led comprehensive data science project development across multiple industries.

Co-founded company and supervised end-to-end development of data science projects
Led foundation of "The Click Reader" ed-tech platform with 10,000+ active learners
Produced 20+ data science courses covering diverse domains and applications
Developed solutions for tourism, healthcare (EEG data), and nuclear research (CERN)

Docsumo

Associate Data Scientist

June 2018 – Feb 2019

Specialized in computer vision and document analysis, developing intelligent document processing solutions.

Developed image segmentation models for document analysis using Mask-RCNN and YOLO3
Achieved 80% accuracy in processing invoices, bank cheques, and ID cards
Optimized data preprocessing pipelines, reducing errors by 10% and training time by 30%
Mentored interns in Python, data analysis, and machine learning fundamentals

Impact: 80% accuracy with 30% faster training and 10% error reduction

Technical Skills

Comprehensive expertise across the full data science stack, from data engineering to model deployment.

Programming & Frameworks

Python R SQL PySpark Apache Spark Flask Django NoSQL

Machine Learning & AI

TensorFlow PyTorch Scikit-learn Transformers HuggingFace OpenAI API LangChain BERT/GPT Models

Computer Vision & NLP

OpenCV YOLO Mask-RCNN SpaCy NLTK Sentence Transformers Named Entity Recognition Text Classification

Data Visualization

Plotly Dash Streamlit PowerBI Matplotlib Seaborn Bokeh D3.js

Cloud & Big Data

Databricks AWS SageMaker Azure ML Google Cloud AI Snowflake Apache Kafka Delta Lake Data Factory

MLOps & DevOps

MLFlow Apache Airflow Docker Kubernetes Git/GitHub Azure DevOps CI/CD Pipelines Model Monitoring

Featured Projects

A selection of impactful projects demonstrating expertise in ML, NLP, and data-driven solutions.

Aircraft Maintenance Intelligence System

Developed an NLP-based application for predictive maintenance alerts using semantic similarity models, knowledge graphs, and automated workflow systems for aerospace components analysis.

Reduced operational downtime by 80% and successfully onboarded first customer

Reinforcement Learning for Epidemic Control

Research project applying reinforcement learning techniques for evaluation and decision-support of lockdown policies during epidemic outbreaks, contributing to public health policy optimization.

Published at IISE Annual Conference 2023 & MIE Showcase Winner

ML-Powered Admissions System

Led development of end-to-end machine learning system for university graduate admissions, incorporating predictive modeling, automated decision workflows, and performance monitoring.

Achieved 75-80% prediction accuracy with 15-member team leadership

Medical Image Analysis for Melanoma Detection

Developed deep learning solution for melanoma detection from medical images using advanced computer vision techniques, contributing to early cancer detection capabilities.

Top 1% in Kaggle SIIM-ISIC Competition (Silver Medal)

Intelligent Document Processing

Built image segmentation models for automated document analysis using Mask-RCNN and YOLO3, processing invoices, bank cheques, and identification cards with high accuracy.

80% accuracy with 30% reduction in training time and 10% error reduction

Real Estate Price Optimization

Applied advanced machine learning techniques including ensemble methods and time-series forecasting for home rental price predictions and market analysis optimization.

Generated 5% sales increase through targeted marketing strategies

Credit Card Fraud Detection System

Developed an advanced fraud detection system using Autoencoders for anomaly detection, capable of identifying fraudulent transactions in real-time with high precision and recall.

Enhanced financial security through automated fraud prevention

Nepali Language NLP with BERT

Built offensive language detection system for Nepali text using BERT transformers, addressing the unique challenges of low-resource language processing and cultural context.

Advanced NLP capabilities for underrepresented languages

COVID-19 Mask Detection in Crowds

Developed deep learning solution using Keras for real-time mask detection in crowded environments, contributing to public health safety measures during the pandemic.

Supported public health initiatives during COVID-19 pandemic

E-commerce Customer Segmentation

Applied clustering algorithms and behavioral analysis to segment e-commerce customers, enabling personalized marketing strategies and improved customer retention rates.

Enhanced customer targeting and personalization strategies

The Click Reader EdTech Platform

Co-founded and developed comprehensive online learning platform with 10,000+ active learners, creating 20+ data science courses covering machine learning, analytics, and programming.

Educated 10,000+ learners globally in data science

5+

6