Sai Chaitanya Pachipulusu

About Sai Chaitanya

Experience

1. C

   CORtracker 360

   • Data Analyst

     Jun 2025 - Present · 3 mos·United States·Current

     • Developed scalable data pipelines and reporting solutions with APIs, Tableau/PowerBI interactive dashboards, and Python (Pandas, NumPy) automation, cutting reporting timelines by 86% and enabling real-time analytics in high-volume environments • Performed statistical analyses with K-means clustering, seasonality modeling with time-series techniques (ARIMA), and A/B testing in Python (Scikit-learn, Statsmodels), increasing creator participation by 13% and optimizing engagement strategies • Created Tableau/PowerBI dashboards with ML insights (XGBoost) reducing reporting time 35%; maintained KPI dashboards via SQL/Python automation for metric tracking and decisions in Snowflake • Optimized ETL pipelines using SQL, Snowflake warehousing, Airflow orchestration, and ML anomaly detection (Autoencoders), improving data quality 30% for cloud initiatives in AWS Redshift • Partnered with 3 engineers to define requirements and build SQL/DBT pipelines with PySpark, boosting transformation speed 30%; developed SQL/PySpark/DBT workflows with ML preprocessing (Pandas/NumPy) for AWS Redshift analytics • Led A/B testing frameworks with SQL and Python (Statsmodels, SciPy) to evaluate feature impacts, resulting in 15% uplift in user retention; mentored junior analysts on experimental design in Agile environments

2. C

   Community Dreams Foundation · Full-time

   • Machine Learning Engineer

     Jul 2024 - Jun 2025 · 1 yr·United States

     ◦ Architected an HR tool using BERT, RoBERTa, and Sentence-BERT embeddings to match resumes with job descriptions, cutting manual screening by 89% from 10 hours to 1 hour per job opening and speeding up the hiring by 40% through context-aware ranking ◦ Built a cloud-native pipeline with Python, FastAPI, and Kubeflow on Kubernetes for automated interview scheduling, achieving 92% candidate selection precision measured in pilot with 50 companies ◦ Automated rejection emails with sentiment-aware templates (VADER score >0.6), handling 200+ weekly communications and reducing admin work by 90% from 15 hours/week to 1.5 hours/week while ensuring empathetic, bias-free communication ◦ Automated the process of sending personalized rejection emails, reducing time spent by 90% ◦ Designed a RAG chatbot with Mistral on AWS SageMaker, achieving 90% response relevance as measured by BLEU and ROUGE-L scores against human answers on a benchmark set of 500 historical support cases, reducing ticket escalations by 25%

3. C

   CGI · Full-time

   • Associate Software Engineer

     Sep 2020 - Jun 2022 · 1 yr 10 mos·Bengaluru, Karnataka, India

     ◦ Developed a real-time ingest layer using Kafka Connect to capture 8 sensor data streams (400 events/second) from factory equipment, reducing data availability lag from overnight batch to <5 minutes for maintenance teams ◦ Wrote and maintained 15+ Python ETL scripts to process daily Shell refinery data files (CSV, JSON) into a centralized SQL data warehouse, enabling dashboard KPIs previously unavailable to operating teams ◦ Led migration of 52 legacy servers to AWS EC2 (t3.xlarge instances) using Terraform, reducing monthly costs by $18k and ensuring 99.9% uptime over 6-month period ◦ Redesigned Databricks medallion architecture by implementing automatic schema validation, data quality checks, and incremental processing patterns, reducing job failures from 12 weekly incidents to 3 (75% decrease) and cutting average recovery time from 4 hours to 45 minutes for Shell’s refinery sensor data lake ◦ Achievement: Awarded Best Employee for Q4 2021 for exceptional contributions to project efficiency and innovation, leading 30% of the overall data migration effort

4. I

   ImbueDesk ENS Pvt Ltd · Full-time

   • Machine Leanring Engineer

     May 2018 - Aug 2020 · 2 yrs 4 mos·Hyderabad, Telangana, India · On-site

     ◦ Developed facial expression recognition system using OpenCV, TensorFlow achieving 97% accuracy on FER2013 dataset ◦ Designed an image processing pipeline with Tesseract OCR for vehicle ID recognition (50k plates/day), orchestrated with Kubernetes ◦ Created and deployed predictive maintenance dashboards with Python visualization tools on AWS Beanstalk that reduced equipment downtime by 28% ◦ Built a Kafka-based image processing pipeline processing 35MB/hour, reducing processing latency from 1.2s to 0.4s per image with a 3-node consumer topology. Implemented back-pressure handling for peak traffic periods (7AM-9AM) when processing volume increased by 300%

Education

1. 
   Stevens Institute of TechnologyMaster's degree, Machine LearningSep 2022 - May 2024

   Grade: 3.9

2. 
   Sreenidhi Institute of Science and TechnologyBachelor of Technology, Information Technology2016 - 2020

Skills