Kashyap Balakavi

Kashyap Balakavi

AI/ML Engineer • Researcher • Applied Innovations

GitHub LinkedIn Email Resume
AI/ML Engineer passionate about transforming bold ideas into data-powered impact.
Bridging research and real-world deployment, across health, climate, and business, by designing scalable solutions that turn complex information into actionable intelligence.
Skilled at leading research-to-production in computer vision, deep learning, NLP, and geospatial systems. If it’s AI and it matters, I want to build it.
Turning datasets into decisions that shape the future.
3+ Years
Experience
15+
ML Models Built
4
Papers Published
12%
Accuracy Lift

About Me

I build intelligent systems that matter: I blend deep learning, computer vision, NLP and scalable cloud systems with a research-first engineering approach. I focus on end-to-end solutions, from model development to production deployment and visualization.
  • Open to: AI, ML, DS & R&D roles.
  • Research Focus: Deep Learning · Computer Vision · Geospatial AI · Visualization.
  • Strengths: Scalable ML, Model Deployment, Visualization, Cross-functional leadership.
  • Tech: Python · TensorFlow · PyTorch · SQL · Cloud · Docker · Kubernetes.

Capabilities

Core Expertise

  • ML/DL: TensorFlow, PyTorch, scikit-learn, XGBoost.
  • MLOps: Docker, Kubernetes, MLflow, Airflow.
  • Cloud: AWS, Azure, Google Earth Engine.
  • Data: SQL, Spark, Snowflake, BigQuery, ETL.

Specializations

  • Computer Vision (CNNs, U-Net, GANs).
  • NLP & Transformers (BERT).
  • Time-Series & Forecasting.
  • Geospatial AI & Remote Sensing.

Impact Delivered

  • 12% accuracy improvement in wildfire models.
  • 40% reduction in data processing time.
  • 19% reduction in donor churn via predictive analytics.
  • 10% ROI lift through A/B testing.

Professional Profile

  • Education: MS Data Science, UAB (GPA: 3.9).
  • Location: Huntsville, AL, USA (Open to relocation).
  • Open to: Data Scientist, AI/ML Engineer, Data Analyst, Research roles.
  • Work Style: Remote, Hybrid, or On-site.

What Sets Me Apart

  • End-to-end ML: From data pipelines to production deployment.
  • Research + Industry: Published papers & real-world systems.
  • Cross-domain: Healthcare, climate, IT, business analytics.
  • Stakeholder communication: Translating ML to business value.

Expertise

Professional Experience

Virginia Tech Research Assistant, AI/ML

Aug 2025 – Present
  • ML pipelines for wildfire pollution, improved prediction by 12%.
  • Drought/fire risk models using 40-year climate datasets.
  • Automated satellite pre-processing with Earth Engine.

University of Alabama Huntsville Research Associate, AI/ML (ESSC)

Jul 2024 – Aug 2025
  • Deep learning for wildfire imagery (TensorFlow, U-Net, GRU).
  • Cloud workflows for NASA teams.
  • Deployed GAN for burn area mapping.

USRA Data Analyst, Wildfire/Deforestation

May 2023 – Jul 2024
  • Boosted wildfire and deforestation detection using CNN-RNN on satellite imagery.
  • Integrated ML results into NASA systems for faster decisions.
  • Enhanced pipelines for rapid disaster response.

University of Alabama at Birmingham Graduate Assistant

May 2022 – Apr 2023
  • Created intuitive data visualizations of virology datasets with D3.js.
  • Added interactive search and lineage tools for fast exploration.
  • Supported impactful research and team collaboration.

University of Alabama at Birmingham Annual Giving Intern

Feb 2022 – Apr 2022
  • Managed donor outreach campaigns and contributed to CRM data hygiene.
  • Analyzed donor engagement metrics to improve fundraising communications.
  • Produced reports and collateral used by the development team.

TCS Assistant Systems Engineer Trainee

Jan 2021 – Dec 2021
  • Built a scalable inventory management system using React, Fluent UI, and SharePoint.
  • Automated business workflows with Power Automate, reducing delays by 21%.
  • Delivered real-time dashboards in Power BI to accelerate data-driven leadership decisions.

Featured Projects

Visual Taxonomy Browser thumbnail

Visual Taxonomy Browser [ICTV Official]

Lead Developer · Paper Author

Interactive D3.js visual analytics for exploring global virus taxonomy. Built for ICTV to make taxonomy navigation intuitive for researchers and the public.

D3.jsVisualizationJavaScript
Wildfire mapping thumbnail

Wildfire & Burnt Area Mapping

CNN-RNN, GAN, GEE

Deep learning models and Earth Engine pipelines for burnt-area mapping and rapid disaster response.

TensorFlowEarth EngineGAN
Heart risk thumbnail

Heart Disease Risk Engine

XGBoost, SVM, Flask API, AWS

Predictive pipeline and API for cardiometabolic risk scoring, with feature explainability and deployment on cloud services.

XGBoostFlaskAWS
Product recommender thumbnail

AI Product Recommender

BERT, FastAPI, Azure

Hybrid NLP and collaborative filtering recommender serving 20K+ items with a production API.

BERTFastAPIAzure

Selected Publications

Visual Taxonomy Browser for Virus Classification [ICTV Live Tool]

Mapping Burnt Areas via Deep Learning & Satellite Imagery

Peer-reviewed in IEEE VIS, PLOS ONE, Sensors, and CRC Press. Full Google Scholar →.

Education

MS, Data Science

Univ. of Alabama at Birmingham (2022–2023)
  • GPA: 3.9 | Focus: Algorithms, Deep Learning, Systems Programming.

Contact