Data Science & Machine Learning

Turning messy, real-world data into decisions worth acting on.

M.S. Data Science graduate building end-to-end ML systems — from early risk detection to conflict forecasting to power-efficient hardware. Based in Denver, CO. Open to Data Science, ML, Analyst, and Data Engineer roles.

Denver, CO chamukottur@gmail.com (720) 839-6576
risk_signal.monitor — case linkage model live readout
input: spatial · temporal · behavioral confidence: rising
Get in touch GitHub ↗ LinkedIn ↗
// 01 Summary

Recent Master of Science in Data Science graduate with a strong foundation in machine learning, statistical analysis, data visualization, and predictive modeling. Experienced in developing end-to-end data science solutions through academic research and real-world projects involving NLP, classification, forecasting, and large-scale data analysis. Proficient in Python, SQL, Tableau, Power BI, and modern machine learning frameworks — and eager to turn complex datasets into decisions that hold up outside the notebook.

// 02 Skills

Languages

PythonSQLR

ML & AI

Machine LearningDeep LearningComputer VisionCNNsYOLOReinforcement LearningStatistical Modeling

Frameworks

PyTorchTensorFlowOpenCVNumPyPandas

Data Processing

PreprocessingDataset MgmtAugmentationFeature EngineeringETL

Visualization

Power BITableauExcel

Databases & Tools

MySQLPostgreSQLMongoDBGitAWSJupyterVS Code
// 03 Projects
Early Risk Assessment & Linkage Analysis of Violent Crime Cases
Data Science Practicum — Chicago Crime Dataset
flags at-risk cases
  • Built an end-to-end ML pipeline to predict case-stalling risk and detect linkages between violent crime incidents, processing large-scale spatial, temporal, and behavioral crime data.
  • Engineered features from raw crime records (location, time patterns, offense behavior) to power risk classification and similarity-based case linkage models.
  • Translated model outputs into an investigative decision-support tool, demonstrating applied ML for public-safety outcomes.
classificationfeature engineeringspatial-temporalPython
Predicting Geopolitical Escalations & Narrative Change in International Conflict
ML & NLP Research — ACLED, GPR, NYT, Reddit, YouTube data (2015–2025)
AUC 0.871
  • Built and validated a binary classification model on a decade of conflict event data (ACLED + GPR), outperforming traditional baselines.
  • Designed a BERT-based NLP pipeline to extract sentiment and narrative-shift signals from multi-source media as early-warning features.
  • Ran comparative evaluation across stable vs. volatile periods, identifying where multimodal signals add the most predictive value.
BERT / NLPbinary classificationmultimodal signalstime series
Power-Efficient Approximate Multiplier for Neural Network Applications
Hardware / ML Systems Research — VLSI Architecture
−15% power
  • Co-designed an 8-bit approximate multiplier architecture to cut power and area cost of neural network inference hardware.
  • Built two novel 4:2 approximate compressors with self-offsetting error, minimizing net approximation error while maximizing efficiency.
  • Validated results on image processing and neural network workloads — published at VDAT 2023.
VLSIhardware-aware MLpublished research
// 04 Education & Certifications

M.S. Data Science

Regis University — May 2026

Dean's Scholarship, 2024–2026

B.Tech, Electronics & Communication Engineering

BVRIT — Jul 2024

Foundation in hardware systems and signal processing.

AI and Emerging Technologies Professional Certification — IIT Hyderabad
Published Research: Power Efficient Approximate Multiplier for NN ApplicationsVDAT 2023
Research Showcase, PresenterSpring 2024
Google Data Analytics Professional Certificate