Full Stack AI Engineer - RAG, Kafka, Azure

Sruthila Bokka

Charlotte, NC +1-980-***-**** ************@*****.*** Linkedin

PROFESSIONAL SUMMARY

Full Stack Engineer with 4+ years of experience architecting AI/ML, Generative AI, and cloud-native systems in healthcare and financial services. Proven expertise building RAG pipelines, LLM-powered applications, and NLP solutions using Python, LangChain, Azure OpenAI, and FAISS, alongside scalable REST/GraphQL APIs with FastAPI, Java, and Spring Boot. Skilled in event-driven architecture (Apache Kafka), containerized deployments (Docker, Kubernetes), and MLOps practices across Azure and AWS. Experienced partnering with clinical and business stakeholders to translate ambiguous requirements into reliable systems, driving measurable gains in latency, retrieval accuracy, and operational efficiency.

EXPERIENCE

Cigna Healthcare Charlotte, NC

Full Stack AI Engineer Oct 2024 – Present

• Built a FHIR-compliant REST API in FastAPI supporting member data exchange with hospital partners, handling claim lookups and eligibility checks at the service layer.

• Reduced diagnostic report turnaround time by extending the API with a RAG-based natural language query layer, using Azure OpenAI and vector search for conversational access to clinical member history.

• Implemented prompt-tuning and few-shot examples in the RAG query layer to improve retrieval accuracy on ambiguous member queries, partnering with a clinical SME to validate outputs before rollout.

• Developed Angular claim submission forms and integrated a document-extraction step (LangChain-based) into the Python backend to auto-populate structured fields from uploaded PDFs, reducing manual entry effort by roughly 30%.

• Built an event-driven claims routing service on Apache Kafka, incorporating an early-stage LLM-based scoring step to flag anomalous claims for the fraud review team.

• Developed a lightweight anomaly-detection model (scikit-learn) on historical claims data to flag likely duplicate submissions before they reached the review queue, cutting 20% of manual triage volume after iterating on false-positive tuning.

• Built ETL pipelines on Azure Data Factory and Databricks (PySpark) for clinical data processing, following internal data-handling and compliance requirements.

• Containerized services with Docker and deployed to AKS, maintaining latency and monitoring (Datadog) within target SLAs.

Piper Sandler Charlotte, NC

Software Developer Feb 2024 – Sep 2024

• Designed a Python-based equity data aggregation service using Apache Kafka for event-driven ingestion, supporting real-time processing across 12 asset classes and reducing data latency by ~40% compared to the prior batch-based pipeline.

• Built interactive financial analytics dashboards with React, TypeScript, and D3.js, streamlining report generation for a team of 150+ equity analysts and cutting manual reporting time by roughly 60%.

• Developed REST and GraphQL APIs (Node.js, SQL Server) to integrate Bloomberg Terminal and Refinitiv market data feeds, providing traders across 4 desks with more consistent, near-real-time access to pricing data.

• Set up CI/CD pipelines with Jenkins and GitHub Actions, introducing blue-green deployments that shortened release cycles from roughly 2 weeks to about 1 week and reduced deployment-related incidents. Dell Technologies India

Software Developer Jul 2020 – Aug 2022

• Developed core modules of an enterprise inventory management platform (Java/Spring Boot, Angular) used to track hardware orders across multiple global fulfillment regions for a large enterprise client base.

• Migrated several legacy monolithic applications to AWS-hosted microservices using Docker and REST APIs, improving scalability and reducing infrastructure overhead.

• Built automated test suites with Selenium and JUnit, expanding critical-path coverage and shortening QA validation time by several days per sprint.

• Optimized SQL Server queries and stored procedures, and introduced Redis caching for high-frequency inventory reads, cutting average API response times significantly.

• Implemented OAuth 2.0 / JWT-based authentication with Spring Security to support secure single sign-on for internal users, and maintained source control workflows in Azure DevOps. ACADEMIC PROJECTS

LLM-Powered Clinical Decision Support System LangChain, OpenAI GPT-4, FAISS, FastAPI, Python

• Built a Retrieval-Augmented Generation (RAG) pipeline using LangChain and GPT-4 over a corpus of ~10,000 clinical guideline documents indexed in FAISS, with a FastAPI backend returning responses in under a second for most queries.

• Experimented with prompt engineering techniques (chain-of-thought, few-shot prompting) to reduce hallucinated answers compared to a zero-shot GPT-4 baseline, evaluating results on a small held-out sample of test questions.

Real-Time Financial Sentiment Analysis Engine BERT, PyTorch, Apache Kafka, AWS Lambda, Python

• Fine-tuned a BERT model on a public financial news sentiment dataset (~200K labeled samples) using PyTorch, and built a Kafka streaming pipeline to simulate real-time inference via AWS Lambda.

• Used MLflow to track experiments across learning rate, batch size, and tokenization variations, improving F1 score from roughly 0.79 to 0.91 over the course of iteration. Scalable Movie Recommendation System PySpark, MLlib, Databricks, AWS S3, Delta Lake

• Built an ALS collaborative filtering model on the MovieLens 10M dataset using PySpark MLlib on Databricks, achieving a mean absolute error of ~0.72, comparable to published benchmarks on the same dataset.

• Set up a Delta Lake-backed pipeline on AWS S3 to simulate a low-latency serving layer for generating recommendations.

TECHNICAL SKILLS

Languages: Python, Java, SQL, TypeScript

Frontend: Angular, React, D3.js, HTML/CSS

Backend & APIs: FastAPI, Spring Boot, Node.js, Express.js, GraphQL, REST APIs, Microservices Architecture, Event-Driven Architecture, OAuth 2.0, JWT, Spring Security, SSO Databases: SQL Server, T-SQL, PostgreSQL, MongoDB, Redis, Apache Cassandra Cloud & DevOps: Microsoft Azure, AWS, Kubernetes, Azure DevOps (TFS), Docker, Terraform, Helm, CI/CD, Jenkins, GitHub Actions, Datadog, Lambda, S3, Bedrock, DynamoDB, Agentcore Data & Messaging: Apache Kafka, PySpark, Databricks, Azure Data Factory, Delta Lake, Snowflake AI/ML & Generative AI: Large Language Models (LLMs), Retrieval-Augmented Generation (RAG), Generative AI, LangChain, Azure OpenAI, OpenAI API, Hugging Face Transformers, Scikit-learn, TensorFlow, PyTorch, Natural Language Processing (NLP), MLflow, Prompt Engineering, Pinecone, FAISS Practices & Tools: System Design, Agile/Scrum, A/B Testing, SonarQube, GitFlow, Git, JIRA, Confluence EDUCATION

Belhaven University Jackson, MS

Doctor of Business Administration in Applied Artificial Intelligence Aug 2026 – Present Boston University Boston, MA

Master of Science in Software Development Sep 2022 – Jan 2024 Gokaraju Rangaraju Institute of Engineering & Technology India Bachelor of Technology, Computer Science July 2018 – May 2022

Contact this candidate