Open to opportunities

ramya sudireddy

@ramyasudireddy

Message

Senior AI/ML and Data Engineer building scalable NLP, fraud, and real-time ML systems.

United States

Message

What I'm looking for

I’m looking for a role where I can build production-ready AI/ML and real-time data platforms with strong MLOps, model governance, and explainability—while mentoring teams to ship reliable, scalable systems.

I’m an AI/ML Engineer and Data Engineer with 9+ years building scalable machine learning, NLP, and real-time data solutions across insurance, financial services, and telecommunications. I focus on production reliability—designing systems that perform in real environments, not just in experiments.

In my current role, I architected a RAG pipeline using LangChain, GPT-4, and Pinecone to power an internal product recommendation assistant, reducing customer support escalations by 28%. I also migrated batch scoring to real-time inference with AWS SageMaker and Kafka, cutting prediction latency from 4 seconds to under 300ms during peak retail events, and led a team of 5 to improve demand forecasting inventory replenishment accuracy by 18% across 1,800+ stores.

I bring end-to-end expertise across ETL, streaming, and MLOps: I’ve established MLflow CI/CD pipelines, centralized feature stores (Tecton + Databricks), and LLM fine-tuning with LoRA/PEFT on Hugging Face for retail personalization. Across domains, I’ve delivered fraud detection, document intelligence, NLP summarization, and computer-vision compliance workflows while applying explainability (SHAP/LIME) and strong governance so models stay trustworthy, measurable, and maintainable.

Experience

Work history, roles, and key accomplishments

Current

Senior AI/ML Engineer

Current

Target

Apr 2025 - Present (1 year 2 months)

Architected a RAG recommendation assistant using LangChain, GPT-4, and Pinecone, reducing customer support escalations by 28%. Migrated batch scoring to real-time inference on AWS SageMaker with Kafka, lowering peak-event prediction latency from 4s to under 300ms.

Lang Chain GPT 4 Pinecone AWS Sagemaker Apache Kafka PyTorch Snowflake YOLOv8 MLFlow fastAPI

ML Engineer

Fidelity Investments

Jun 2024 - Apr 2025 (10 months)

Built predictive risk models (XGBoost, LightGBM, Random Forest) on a 500K+ member dataset to reduce escalation response time by 28%. Designed real-time ingestion with Kafka and Spark Structured Streaming into Azure Data Lake Gen2 and standardized 60+ reusable features across model teams, cutting data prep redundancy by 40%.

XGBoost LightGBM Random Forest Apache Kafka Spark Structured Streaming MLFlow DVC

ML Engineer

Cotiviti

Aug 2023 - May 2024 (9 months)

Developed end-to-end NLP pipelines with spaCy and Hugging Face Transformers to automate clinical text classification, reducing manual review time by 35%. Built scalable ETL for 50M+ records/month using PySpark and deployed real-time inference on AWS SageMaker with Docker/Kubernetes, improving fraud/anomaly detection precision by 22% over legacy rules.

Python SpaCy Pyspark Apache Spark AWS Sagemaker Docker Kubernetes SHAP fastAPI

Data Engineer

American Express

Oct 2020 - Apr 2022 (1 year 6 months)

Productionized a real-time fraud detection engine in Python with Scikit-learn and XGBoost, processing 2M+ daily transactions and reducing false positives by 22%. Built AWS EMR Spark ingestion pipelines into Redshift and implemented Kafka/Kinesis streaming to enable sub-second fraud decisioning.

Python scikit learn XGBoost Apache Spark Amazon Redshift Apache Kafka AWS Kinesis DBT Airflow

Data Engineer

Cox Communications

Nov 2018 - Sep 2020 (1 year 10 months)

Built Python-based ETL pipelines (SQL, Pandas) for network telemetry supporting analytics for 1M+ subscribers. Developed predictive maintenance models with Scikit-learn and Random Forest, reducing network issue MTTR by 28% and improving ETL reliability using Celery and Redis.

Python SQL Pandas scikit learn Random Forest Celery Redis Flask Tableau AWS RDS

Software Engineer

Apollo Hospitals

Jul 2015 - Oct 2018 (3 years 3 months)

Built patient management modules in Django (appointment scheduling and OPD registration) used daily by 500+ hospital staff. Implemented HL7-compliant integrations to eliminate manual diagnostic re-entry and created billing/invoice automation with Python and ReportLab to reduce claim submission delays.