Yaswanth Sivala
AI/ML Engineer • Kansas City, MO • s************@gmail.com • 573****077 • linkedin.com/••••• • drivetube.ai/•••••
Professional Summary
AI/ML Engineer with 3+ years of experience designing, deploying, and operating production AI systems end-to-end — from data pipelines and serving APIs to LLM applications and MLOps. Experienced building agentic and multi-agent LLM workflows with LangChain and LangGraph, retrieval-augmented systems (RAG), and production services using FastAPI, Docker, Kubernetes, MLflow, and observability tooling.
Technical Skills
Programming Languages: Python,Java
Web Technologies: REST APIs
Frameworks and Libraries: LangChain,Hugging Face,TensorFlow,PyTorch,Scikit-learn,FastAPI,Flask
Databases: Snowflake,IBM Cloud,Render,Hugging Face Spaces,PostgreSQL,MySQL,Supabase,SQLite,SQL
Cloud and DevOps: MLflow,Model Registry,Docker,Kubernetes,Helm,GitHub Actions,CI,CD,Prometheus,Grafana,Evidently AI,AWS,Azure
Testing: pytest,Unit Testing
Data and Analytics: XGBoost,spaCy,Feature Engineering,Classification,Clustering,Time Series Modeling,Anomaly Detection,Apache Spark,Databricks,Airflow,Azure Data Factory,ETL,Data Warehousing
Tools and Methodologies: Git,Agile,JIRA
Skills: Microservices,System Design,Azure DevOps,TDD
Generative AI & LLMs: LangGraph,Prompt Engineering,Tool and Function Calling,Retrieval Augmentation,Embeddings,FAISS,ChromaDB,Pinecone,FastEmbed,OpenAI,Azure OpenAI,Whisper,Llama 3
Work Experience
InfoShare Systems, Inc.
Kansas City, MO
AI/ML Engineer
Feb 2025 – Present
Built agentic LLM systems and production AI services for enterprise document intelligence and analytics, supporting enterprise teams with retrieval-augmented answers and high-throughput inference.
Tech Stack: LangChain, LangGraph, OpenAI, Azure OpenAI, ChromaDB, FastAPI, Docker, Kubernetes, MLflow, GitHub Actions, Prometheus, Grafana, Evidently AI, Azure Data Factory, Databricks, Snowflake
- Architected and delivered an agentic LLM document-intelligence platform using LangChain, LangGraph, and OpenAI/Azure OpenAI; implemented tool-calling agents that reduced manual document lookup time by 40% for enterprise teams.
- Designed and implemented a grounded RAG pipeline with ChromaDB, custom chunking, semantic search, and re-ranking, reducing hallucination rate by ~15% and improving citation accuracy for answers over enterprise datasets.
- Built high-throughput FastAPI microservices serving 100K+ daily inference requests; optimized request handling and scaling to reduce downstream API latency by 15% under production load.
- Standardized the ML lifecycle across 5+ services using MLflow and a model registry; containerized models with Docker and Kubernetes and automated build/test/deploy with GitHub Actions, shortening release cycles from days to hours.
- Instrumented production models with Prometheus, Grafana, and Evidently AI for metrics, observability, and drift detection; implemented alerting and reporting to maintain model quality and surface regressions.
- Designed scalable ETL and data-processing pipelines on Azure Data Factory, Databricks, and Snowflake to support enterprise analytics, improving data-processing efficiency by 25%.
Infotech
India
Data Analyst
Mar 2022 – May 2023
Delivered analytics and fraud-detection systems for banking and financial services, building ETL pipelines and models to support operations and executive reporting.
Tech Stack: Python, Pandas, Scikit-learn, Apache Spark, SQL, Power BI, Tableau, ETL
- Developed anomaly-detection models on banking transaction datasets using Python, Pandas, and Scikit-learn, increasing detection accuracy by 18% across millions of records and reducing financial-fraud exposure.
- Built production ETL pipelines aggregating multi-source financial data, cutting report-generation time by 30% for fraud-detection teams and enabling faster operational responses.
- Engineered feature pipelines and scaled preprocessing using Apache Spark and SQL to support daily batch scoring and large-scale transaction processing.
- Designed and deployed K-Means and RFM customer-segmentation models to support marketing and risk teams, improving campaign targeting efficiency by 22%.
- Developed Power BI and Tableau dashboards to visualize model outputs and KPI trends for executives, operationalizing insights and accelerating decision-making.
- Collaborated with data engineering and business stakeholders to establish data quality checks and monitoring, reducing data-related incidents in fraud pipelines.
Ajio (Reliance Retail)
India
Junior Data Scientist (Intern)
Jun 2021 – Feb 2022
Contributed to e-commerce analytics and product-data pipelines, building predictive models and NLP processing for unstructured product information and customer-facing workflows.
Tech Stack: XGBoost, Scikit-learn, spaCy, Azure Cognitive Services, Microsoft Bot Framework, Python, OCR
- Built and tuned XGBoost and Scikit-learn classifiers, achieving a 92% F1-score on targeted predictive benchmarks used by product teams.
- Developed spaCy NLP pipelines for text extraction, tokenization, and entity recognition over unstructured product data, improving metadata coverage and searchability.
- Integrated Azure Cognitive Services OCR to extract product attributes and implemented Microsoft Bot Framework chatbot workflows, accelerating decision cycles by 35%.
- Performed feature engineering, cross-validation, and hyperparameter tuning to optimize model performance for SKU-level predictions and downstream use cases.
- Prepared model evaluation reports and technical handoffs for engineering teams to support productionization and continued model monitoring.
- Partnered with product and merchandising teams to convert model results into actionable business rules for catalog cleanup and recommendation improvements.
Implemented Azure Cognitive Services
Projects
Enterprise RAG AI Assistant
Tools Used: Python, Flask, LangChain, FAISS, FastEmbed, Groq, Docker, GitHub Actions
- Built and deployed an end-to-end RAG application where users upload PDFs and receive LLM answers grounded in their documents with source citations, powered by FAISS vector search and FastEmbed embeddings.
- Engineered a provider-abstraction layer (IBM watsonx.ai → Groq) to avoid vendor lock-in and isolated per-session state using UUID-scoped FAISS indexes.
- Hardened the app against XSS, implemented a 21-test pytest suite, and automated CI with GitHub Actions on a non-root Docker image for secure continuous delivery.
AI Insurance Voice Agent
Tools Used: Python, Flask, Groq, Whisper, Llama 3, Supabase, Render
- Built and deployed a multilingual AI voice-intake platform that transcribes insurance calls in 99+ languages, summarizes them in English, and auto-classifies them into business lines in real time.
- Migrated the AI stack off IBM Watson to the Groq API, reducing run cost and improving transcription speed and language coverage.
- Deployed on Render with Infrastructure-as-Code (render.yaml), environment-based secrets, and HTTPS to provide a live public endpoint for demonstration and testing.
Education
Missouri University of Science and Technology
M.S., Information Science & Technology • Rolla, MO • Aug 2023 – May 2025
Avanthi Institute of Engineering & Technology
B.Tech, Computer Science & Engineering • India • Aug 2019 – May 2023
Certifications
IBM Data Analyst Professional Specialization — IBM
IBM Guardium Data Protection – Practitioner (Advanced) — IBM
IBM Guardium – Technical Sales (Intermediate) — IBM
IBM Insights Sales Foundation — IBM
Powered by Drivetube · Create your own profile at drivetube.ai