Saketh Saridena
Data Scientist & AI EngineerM.S. from CU Boulder (4.0 GPA). I build LLM-powered platforms, production ML pipelines, and GenAI systems that deliver measurable business impact.
01. About Me
I don't just build models. I build AI systems that move the needle. As a Data Scientist at Applexus Technologies, I architected an AI-powered EDA platform that cut client analysis time by 95% and tripled our team's project capacity. That's the kind of impact I chase: measurable, scalable, and impossible to ignore.
I hold an M.S. in Data Science from CU Boulder with a perfect 4.0 GPA, a published IEEE paper, a filed Indian patent, and research presented at PVSEC-31. My toolkit spans the full AI stack, from fine-tuning GPT-4, LLaMA, and Mistral to deploying production pipelines on Azure Databricks and AWS. I've evaluated LLMs for Honda Research Institute, engineered RAG architectures with LangChain, and shipped deep learning models that serve real users.
What sets me apart? I think in business outcomes, not just accuracy scores. Every model I build comes with a clear ROI story, whether that's a 60% reduction in manual effort, a 40% improvement in partner identification, or a predictive maintenance system that scored 0.988 AUC in competition. I bridge the gap between cutting-edge research and products that actually ship.
class SakethSaridena:
def __init__(self):
self.role = "Data Scientist & AI Engineer"
self.location = "Seattle, WA"
self.gpa = 4.0 # Perfect
self.patent = True
self.publications = 3
def impact(self):
return {
"eda_time_saved": "95%",
"capacity_boost": "3x",
"best_auc": 0.988,
}
def philosophy(self):
return "Ship AI that moves the needle"02. Experience
Data Science Consultant
Applexus Technologies · Federal Way, WA- Engineered AI-powered EDA platform using Python, Node.js, and Electron, integrating 50+ statistical tests and local LLM interpretation, reducing client EDA cycle time by 95%
- Assessed AI/ML feasibility for 10+ enterprise use cases, performing feature engineering and ML readiness evaluation, reducing project scoping time by 80%
- Developed spend analytics and demand forecasting frameworks using Azure Databricks & SAP Analytics Cloud for enterprise procurement optimization
Data Analysis Intern
Eco Servants Project · Remote, USA- Developed Sponsorship Data Analysis Framework that increased partner identification by 40% and improved donor targeting accuracy by 35%
- Built Blog Topic and Audience Research Report using SEO and Google Analytics, improving audience reach by 50%
Data Scientist (Capstone)
Honda Research Institute USA · Remote- Built AI-powered data preprocessing pipeline using GPT-4 & MLOps, reducing manual effort by 60%
- Evaluated LLMs (GPT-4, Mistral 7B, Falcon 7B, LLaMA 3) and improved processing accuracy by 30%
- Designed interactive visualization system with Groq's Mixtral-8x7b, achieving 50% improvement in dashboard efficiency
Data Scientist
Pucho Digital Health Inc. · Remote, India- Redesigned ResNet50 architecture, achieving 4% increase in diagnostic precision and 15% boost in COVID-19 detection
- Implemented NLP-based noise reduction using spectral subtraction with MFCC, reaching 85% production success rate
Data Science Research Intern
AI & Robotics Center, VIT-AP University · India- Authored patent (Ref. 202341042839) for intelligent traffic control system using deep learning and reinforcement learning
- Built UAV-based aquaculture monitoring using YOLOv5 at 84% accuracy; presented at PVSEC-31 with 99.3% prediction accuracy
03. Skills
Languages
ML / DL Frameworks
LLMs & Generative AI
Specializations
Data & Cloud
BI & Analytics
04. Featured Projects
EZ Review: Amazon Sentiment Analyzer
Full-stack NLP platform that scrapes Amazon reviews, performs sentiment analysis using three model pipelines (TextBlob, BERT, VADER+BERT hybrid), and generates GPT-powered review summaries for rapid product evaluation.
DASSA Predictive Maintenance
Hackathon-winning predictive maintenance system using 18 sensor readings. SMOTE-balanced pipeline with GridSearchCV-tuned Random Forest and XGBoost, including business impact cost analysis.
WTI Oil Price Forecasting
Time series forecasting of crude oil prices using 6 models including LSTM, LightGBM, and Random Forest. Features lag engineering, rolling statistics, and 3-month recursive forward prediction.
Recipe Recommendation System
Three-tier recommendation engine built on 20,000+ recipes with 680 features. Cosine similarity models with nutrition filtering, dietary preferences, and personalized ingredient weighting.
Renewable Energy Forecaster
SVM regressor with feature engineering to forecast U.S. solar, wind, and hydro energy production through 2034. Improved prediction accuracy by 20% and reduced RMSE by 8%.
Intelligent Traffic Control System
Patented system using deep learning and reinforcement learning to optimize traffic flow and prioritize emergency vehicles. Indian Patent Office (Ref. 202341042839).
05. Research & Publications
Intelligent Traffic Control System
Deep learning and reinforcement learning-based system for optimizing traffic flow and prioritizing emergency vehicles.
Automated Monitoring System for Healthier Aquaculture Farming
UAV-based monitoring system using YOLOv5 and computer vision for aquaculture health assessment.
AI Algorithms for Perovskite Solar Cells Fabrication
Presented AI research achieving 99.3% prediction accuracy for solar cell fabrication optimization.
06. Education
University of Colorado Boulder
M.S. Data ScienceVellore Institute of Technology
B.Tech Computer Science (Data Analytics)07. Get In Touch
Whether you have a challenging data problem, a collaboration idea, or just want to talk AI, I'd love to hear from you. I'm always excited to explore new ways to create impact with machine learning and intelligent systems.