Open to Exploring New Opportunities

Manasa
Dontireddy

Data Scientist & Research Engineer with 6+ years across autonomous vehicles & supply chain engineering. Currently MS Advanced Data Analytics @ UNT · Research Assistant studying LLM financial decision-making.

View My Work ⭐ GitHub Get In Touch
6+
Years Industry Experience
Tata Elxsi → Cognizant
3.9
Graduate GPA at UNT
Advanced Data Analytics MS
2
Research Papers In Press
CS Procedia · ACM Transactions
L5
Autonomous Driving Validated
ZOOX · Las Vegas & San Francisco
01

About Me

💡
Who I Am
I'm a data scientist and research engineer with 6+ years across autonomous vehicle systems (Tata Elxsi / ZOOX), global supply-chain engineering (Cognizant / Walmart), and academic research in quantitative AI finance. My career bridges production engineering with modern machine learning.
🤖
Current Research @ UNT
At the University of North Texas, I'm a Research Assistant under Prof. Erdem Orhan, running a live head-to-head experiment where four leading LLMs (Claude, ChatGPT, DeepSeek, Grok) each manage $1,000 in virtual US equity capital — tracked daily against the S&P 500.
✈️
American Airlines Capstone — Flight Delay Prediction
For my capstone project, I partnered with American Airlines to build a flight arrival delay prediction model — applying machine learning techniques to real operational data including weather, routing, and scheduling factors to translate complex datasets into actionable on-time performance insights.
📄
Research Interests & Publications
My research interests span predictive modeling, deep learning, time series forecasting, and AI for decision-making under uncertainty. I'm open to exploring new opportunities in data science, machine learning, and AI research. I've contributed to two peer-reviewed papers (in press) applying Q-learning hybrid metaheuristics to combinatorial optimization problems, and presented a research poster at UNT on LLM financial performance evaluation.
Programming
PythonCC++SQLPro-CShell Scripting
Machine Learning
PyTorchscikit-learnCNNsARIMA/SARIMARegressionSVMRandom ForestsNeural Networks
Analytics & Visualization
TableauPower BIMatplotlibEDAHypothesis TestingFeature Engineering
CV & Autonomous Systems
OpenCVImage ProcessingADASKalman FilterLevel-5 AVLinux/Unix
Cloud & Dev
GCPGitHubJupyterVS CodeAgile/ScrumJIRACI/CD
🎓 GPA 3.9 / 4.0
📄 2 Papers In Press
🌐 Open to Opportunities
🤝 Research Collaborations Welcome
02

Professional Experience

🎓 University of North Texas — Research Assistant
Dec 2025 – Present
Supervised by Prof. Erdem Orhan · Denton, Texas
📍 Current Position
LLMsPythonFinancial ModelingPortfolio SimulationAPI Data Retrieval (yfinance)Time-Series AnalysisPlotly Data VisualizationQuantitative Finance / Risk AnalysisS&P 500Risk AnalyticsWeb DashboardsResearch Poster
🌐 Live Dashboard →
Cognizant Technology Solutions (CTS) — Walmart
Jul 2022 – Aug 2024
Software Engineer – Global Logistics Systems (Supply Chain)
📍 Hyderabad, India · 2 Years
  • Worked on Walmart's Global Logistics Systems (GLS) supporting distribution center (DC) and supply chain operations across international markets including China and Canada.
  • Designed and implemented a mapping module for the China market using Pro-C, SQL, Python scripting, and Shell scripting to enable efficient discovery of item-level information at the distribution center level, improving data accessibility and operational efficiency.
  • Led development and deployment of SAMS Holding Company DC modules covering the end-to-end GLS workflow from receiving to invoicing, implemented using 4GL and Citrix UI applications.
  • Developed and deployed the Market Fulfillment Center (MFC) system for the Canada market using Citrix-based applications, completing full module implementation in under 4 months.
  • Monitored global operational defects and system issues across distribution centers by analyzing logs and system data, identifying root causes and implementing prioritized fixes — reducing recurring issues by ~8%.
  • Collaborated with cross-functional engineering and operations teams to resolve system defects, improving overall supply chain system reliability.
  • Supported backend logistics systems and workflow automation, ensuring seamless coordination between DC processes and enterprise systems.
Technical Skills Used
Programming:Python, SQL, C, Pro-C, Shell Scripting
Systems & Platforms:4GL, Citrix UI, Linux/Unix
Domain:Supply Chain, Distribution Center Ops, Global Logistics Systems (GLS), MFC
SQLPythonPro-C4GLCitrix UIShell ScriptingLinux/UnixRoot Cause Analysis
Tata Elxsi — Client: ZOOX (Amazon)
Nov 2017 – Jun 2022
Research Engineer – Autonomous Vehicles
📍 Bangalore, India · 4+ Years
  • Designed, developed, and deployed Advanced Driver Assistance Systems (ADAS) and Level-5 autonomous driving algorithms using C++ and Python, focusing on scalable and predictive system behavior for major OEMs and Tier-1 suppliers.
  • Tested and validated Level-5 autonomous driving software for ZOOX vehicles across multiple real-world driving scenarios using virtual simulation platforms representing road environments in Las Vegas and San Francisco.
  • Developed and validated lane detection algorithms using computer vision and machine learning techniques, evaluated on large-scale datasets and simulation environments to improve model accuracy and robustness.
  • Implemented Kalman Filter–based tracking algorithms to improve object and lane tracking stability across dynamic driving conditions.
  • Performed data analysis and log debugging to identify root causes of autonomous vehicle disengagements (automatic → manual mode), categorizing failures into structured buckets and sub-buckets to enhance system reliability and safety.
  • Conducted testing, validation, and performance evaluation of perception modules across diverse environmental conditions and edge cases.
  • Contributed to continuous integration and system integration pipelines, ensuring stable and maintainable development environments for autonomous driving modules.
  • Collaborated with cross-functional teams (perception, planning, simulation) and mentored junior engineers to foster a collaborative development environment.
  • Took ownership of ZOOX perception buckets including Traffic Light Junction, Perception Root Cause Analysis (PCP-RCA), Zoox Road Network (ZRN), and Narrow Handling, reviewing and handling complex perception discrepancies to drive system improvements.
Technical Skills Used
Programming:C++, Python, SQL, Shell Scripting
Autonomous Systems & ML:ADAS, Level-5 AV, Predictive Modeling, Algorithm Development
Computer Vision:Lane Detection, Object Tracking, Image Processing, Kalman Filters
Tools:Linux, Virtual Simulation Platforms, CI/CD Pipelines, Git
C++PythonComputer VisionKalman FilterADASLevel-5 AVLinuxCI/CDPredictive Modeling
03

Featured Projects

Model Accuracy Progression · CIFAR-10 ~90% FCN CNN +pool +aug +reg AdamW CosAnn StepLR
🧠
Deep Learning for CIFAR-10 Classification

Progressed from FCN baseline → CNN with pooling → data augmentation → regularization → AdamW optimizer with Cosine Annealing & StepLR schedulers. Achieved ~90% accuracy and analyzed class-wise performance.

PyTorchCNNAdamWPython
GitHub →
Zillow US Housing · 6-Month Forecast forecast → ARIMA/SARIMA · ADF Stationarity · Top 20 US Regions
🏠
Time Series Forecasting of US Housing Prices

Analyzed Zillow data to forecast 6-month trends across the top 20 US regions. Applied seasonality decomposition and ADF testing for stationarity, comparing ARIMA vs SARIMA for optimal prediction accuracy.

ARIMASARIMATime SeriesPython
GitHub →
Price Distribution by Travel Class Budget Economy Business First
✈️
Statistical Analysis of Airline Ticket Prices

Analyzed Spring 2024 Airline Dataset using EDA charts, hypothesis testing, ANOVA, and linear regression. Visualized pricing variability by airline, travel class, and season to derive actionable insights on price determinants.

ANOVARegressionEDAStatistics
GitHub →
US Industrial CO₂ Emissions 2011–2022 Total Emissions Industrial Sector Tableau · Power BI
🌿
Industrial Carbon Emissions Analysis

Analyzed U.S. Emissions & Industrial Data (2011–2022) leveraging Tableau dashboards and Power BI to visualize emission trends and patterns. Used Tableau Prep and Power BI for data cleaning and insight generation.

TableauPower BITableau PrepData Viz
GitHub →
Safety Rating vs Accident Severity Correlation NHTSA FARS Dataset · Jupyter Notebook · EDA
🚗
NHTSA Accident & Vehicle Safety Analysis

Comprehensive analysis of NHTSA datasets (FARS & vehicle safety data) to uncover correlations between vehicle characteristics, safety ratings, and accident severity across timing, geography, and vehicle type classifications.

PythonPandasJupyterEDA
GitHub →
Database Schema · MySQL Relational Design USERS user_id PK name, email auth_hash ACCOUNTS acct_id PK user_id FK balance, type TRANSACTIONS txn_id PK acct_id FK amount, date MySQL · Python · Secure Banking System
🏦
Virtual Banking System

Secure, user-friendly banking system with CRUD operations for account management, transactions, and authentication. Robust exception handling and a relational MySQL database ensuring data integrity.

PythonMySQLCRUDDatabase Design
GitHub →
04

Research

📋 Research Poster — Presented at UNT
Feb / March 2026
Supervised by Dr. Orhan Erdem · University of North Texas
📍 UNT Research Showcase · Denton, Texas
  • A Live Portfolio Battle: Evaluating Claude AI, ChatGPT, DeepSeek & Grok on Real US Equity Markets — February 2026
  • Each LLM managed $1,000 in virtual US equity capital, tracked daily against the S&P 500 benchmark across 4 weeks.
  • All four LLMs underperformed the S&P 500 in February 2026 — ChatGPT was the clear leader at −3.93% vs Grok's −16.25%, a $123 spread between best and worst.
LLMsPortfolio AnalysisS&P 500PythonYahoo FinanceRisk AnalyticsResearch Poster
🌐 View Live Dashboard →

📄 Full Research Poster

🌐 Live Dashboard →
LLM Portfolio Battle Research Poster — UNT 2026

Manasa Dontireddy & Dr. Orhan Erdem · University of North Texas · February 2026

A Live Portfolio Battle: Evaluating Claude AI, ChatGPT, DeepSeek & Grok on Real US Equity Markets
Manasa Dontireddy & Dr. Orhan Erdem · University of North Texas · February 2026
📌 Poster Presented 🌐 Live Dashboard
4
LLMs Compared
-3.93%
Best (ChatGPT)
-16.25%
Worst (Grok)
$123
Best vs Worst Spread
0/4
LLMs Beat S&P 500
Cumulative Portfolio Value vs S&P 500 — Feb 2026
Live Tracker Active
View Live Dashboard →
$1020 $1000 $980 $960 $940 Feb2 Feb6 Feb10 Feb13 Feb18 Feb23 Feb27 S&P 500 ChatGPT Claude AI DeepSeek Grok

📄 Research Publications

01
A Q-learning based Mayfly Optimization Study on Energy Efficient Multi-Objective Flow Shop Scheduling Problem with Setup TimesIn Press
Vigneshwar Pesaru, Manasa Dontireddy
Computer Science Procedia
Proposes a hybrid Q-learning enhanced Mayfly Optimization algorithm for energy-efficient multi-objective flow shop scheduling with setup times — addressing a critical industrial challenge with direct implications for energy reduction and manufacturing efficiency.
02
A Comparative Optimization Study on Improved Mayfly and Grey-wolf Algorithm with Q-Learning on Energy Efficient Permutation Flow Shop Scheduling ProblemIn Press
Vigneshwar Pesaru, Manasa Dontireddy
ACM Transactions on Algorithms
Benchmarks improved Mayfly and Grey-wolf algorithms augmented with Q-learning for the energy-efficient Permutation Flow Shop Scheduling Problem (PFSP), advancing metaheuristic reinforcement learning hybrids for combinatorial optimization.
05

GitHub Portfolio

Active repositories → github.com/ManasaReddy2417

📊 NHTSA-Accidental-Analysis
Correlation analysis of vehicle safety ratings and accident severity using NHTSA FARS dataset.
Jupyter Notebook
🌿 US-Environmental-Emissions
Tableau and Power BI dashboards analyzing US industrial emissions 2011–2022.
Tableau / Power BI
✈️ Airline-Pricing-Analysis
EDA, ANOVA, and regression modeling on Spring 2024 airline pricing dataset.
Python
🏠 Zillow-House-Price-Forecasting
ARIMA/SARIMA time series forecasting of US housing prices across top 20 regions.
Python
🎓 DataCamp-Certifications
Completed certifications in Python, Data Science, Pandas, EDA, Visualization, and Time Series.
Python
🤖 LLM-Performance-Tracker
Live dashboard tracking LLM portfolio performance vs S&P 500 — active research project.
JavaScript · Live
06

Education

🎓
Master of Science
Advanced Data Analytics
University of North Texas
Aug 2024 – May 2026 GPA 3.9 / 4.0 Denton, TX
Relevant Coursework
Machine LearningDeep LearningData AnalyticsStatistical ModellingTableau / Power BIBig DataAgile for AnalyticsCapstone ProjectData Harvesting & Storage
✈️ Capstone: American Airlines — Flight Arrival Delay Prediction Project
⚙️
Bachelor of Engineering
Computer Science and Engineering
VVIT College of Engineering
Aug 2013 – May 2017 GPA 3.5 / 4.0 Guntur, India
Relevant Coursework
Data StructuresOSDBMSSoftware TestingComputer ArchitectureCryptography & NetworksProbability & StatisticsC / C++ / Python / JavaSystem Design
07

Get In Touch

I'm open to exploring new opportunities in data science, machine learning, AI research, and engineering roles. Happy to connect for research collaborations, full-time positions, or conversations about quantitative AI, NLP, and autonomous systems.

GitHub 💼 LinkedIn 🌐 Live Research

Say Hello 👋

Click the address below to copy it to your clipboard.

Click to copy email address
✅ Copied to clipboard!
Open in Mail App →