Sruthi Reddy Narapareddy Senior ML Engineer @ ZS | Data Science at UIC |

Senior Machine Learning Engineer

ZS Associates

Jan 2025 - Aug 2025 (8 months)

Architected AI-driven solutions using LangGraph to integrate multi-agent LLM workflows for decision-making and automation. Designed workflows incorporating planner, retriever, critique, and validation agents to support dynamic, context-aware task execution. Leading the design and development of prompt generation, authoring, and data ingestion systems leveraging Textract, RAG, LangGraph, and Knowledge Graphs to improve information retrieval, task routing, and memory management. Currently leading a team of three in developing the AI module and deployment. Developed data ingestion, RAG retrieval, and multi-agent orchestration modules. Designed a robust data model incorporating versioning, change tracking, and hierarchical content structures to

Senior AI Engineer

Centific

Apr 2024 - Dec 2024 (9 months)

Evaluated Language Models (LLMs) and Vision Models on tasks such as spell check, harmful content detection, keywords extraction, decontextualization, and generating image descriptions. Built and deployed RAG pipelines to generate answers specific to the client's domain and products. Developed and fine-tuned prompts using techniques like few-shot learning and Chain of Thought to enhance quality and consistency metrics. Built CI/CD pipelines using LLMs to automate quality assurance, reducing project timelines by 30% and costs by 25%, thus streamlining the entire process. Leveraged multiple LLMs to detect harmful content and create red teaming (jailbreak) prompts for assessing model robustness and safety, ensuring adherence to responsible AI s

AI Engineer

ZS Associates

Feb 2021 - Mar 2024 (3 years 2 months)

Built computer vision applications deployed on edge devices to process multiple cameras and perform video analytics (detection, tracking). Trained models on data for fight-detection, people-store movement, dwell time, and tuned to infer any violence in store and customer behavior. Developed ML pipelines and DevOps processes to automate training and registering Docker images for deployment on edge devices and Triton servers. Forecasted sales for 600 CPG products using a Time Series model, achieving a 16% WMAPE score. The model recommended pricing, promotions, and display promotion ads, creating a 25% uplift in sales of products. Engineered new features to capture the impact of display and promotion strategies on sales performance. Constructe

Data Science Intern [Capstone]

Cresco Labs

Aug 2020 - Dec 2020 (5 months)

Multi-variate Time Series Forecasting, Predictive Analytics. Performed exploratory analysis, tests to check stationarity and seasonality in the time series data. Created models with random forest and LSTM to forecast the demand of sales for each product in the store with a 6.3 RMSE score. Quantified the effect of different factors like weekend, discount, and brand on the demand using a linear regression model.

Data Analyst

University of Illinois at Chicago

Oct 2019 - Oct 2020 (1 year 1 month)

Student Affairs Technology | Data Mining, Data Analysis, SQL, Tableau. Designed Data Pipelines to transfer data from multiple sources to Datamart and transform for deeper analysis. Developed interactive Tableau dashboards to assess student participation in events, student employment, and CSRC data. Recommended new strategies to improve student participation based on historical patterns observed in the dashboards. Generated Ad-hoc reports using SQL by connecting to disparate data sources (Microsoft SQL Server).

Applications Developer 2

Oracle Financial Services Software Ltd

Nov 2016 - Jul 2019 (2 years 9 months)

Data Management, Automation, Application Development, Java, SQL, PL/SQL. Using PL/SQL, performed data manipulation on tables in Oracle DB that are seamlessly linked to the front end of a complex banking application. Proactively enhanced the UI of the application using Oracle FLEXCUBE ODT tool to increase user friendliness. Built an Automation framework using Java and Selenium for quality analysis of FCUBS, which reduced testing time by 50%. Developed new features in Oracle Banking Payments (OBPM) Application using Java, SQL, JPA to enhance cross-border payments. Optimized search queries, which reduced the validation time of a payment transaction by 40%.

Sruthi Reddy Narapareddy

About

Detail

Timeline

Résumé