AI/ML and Generative AI Engineer

SUDHEER SUDA

AI/ML Engineer Data Scientist Generative AI, Agentic AI & LLM Specialist

*************@*****.*** 945-***-**** linkedin.com/in/sudasudheer github.com/sudasudheerkumar

PROFESSIONAL SUMMARY

•AI/ML Engineer & Data Scientist with 8+ years of experience designing and deploying Generative AI, Agentic AI, and Large Language Model (LLM) solutions across finance, healthcare, and manufacturing domains, with deep expertise in enterprise data warehousing, time series analysis, and revenue/pricing optimization.

•Expert in transformer-based architectures and production-grade LLM applications, including RAG pipelines, multi-agent systems, and intelligent assistants using GPT-4, LLaMA, and Mistral.

•Hands-on experience with Agentic AI frameworks (LangGraph, CrewAI, LangChain) for autonomous reasoning, multi-agent orchestration, tool-calling, and decision-making workflows.

•Skilled in LLM fine-tuning (LoRA, QLoRA, PEFT), prompt engineering, and vector databases (Pinecone, ChromaDB, Weaviate, FAISS) for semantic search and retrieval-augmented generation.

•Proficient in MLOps practices including CI/CD for ML, experiment tracking and model versioning (MLflow), containerization, and scalable deployment with Docker and Kubernetes.

•Effective communicator and cross-functional collaborator skilled at translating complex AI/ML concepts for business stakeholders and executive leadership; experienced mentoring junior engineers and driving adoption of best practices across teams.

•Experienced in building and integrating RESTful APIs and microservices (FastAPI, Flask, Django REST) for AI services and enterprise system integration.

•Strong background in data engineering and ETL using PySpark, Apache Airflow, Kafka, Dataflow, and BigQuery for large-scale data processing.

•Proficient in Python, TensorFlow, PyTorch, Scikit-learn, and XGBoost for classification, regression, and deep learning models (CNNs, RNNs, LSTMs, Transformers).

•Cloud experience across AWS (SageMaker, Lambda, ECS, Glue, S3, Redshift) and GCP (Vertex AI, BigQuery, Dataflow, Document AI), with infrastructure-as-code via Terraform.

•Solid understanding of Agile/SDLC methodologies and proven success collaborating in cross-functional, onshore/offshore teams.

TECHNICAL SKILLS

•Languages: Python (Expert), SQL

•Generative AI & LLMs: GPT-4, GPT-3.5, LLaMA, Mistral, Gemini, Claude, OpenAI API, AWS Bedrock (Claude, Titan, Llama 2), Fine-Tuning (LoRA, QLoRA, PEFT), Prompt Engineering, RAG Pipelines, LangChain, LangGraph, LlamaIndex, CrewAI, Vector Databases (Pinecone, ChromaDB, Weaviate, FAISS)

•Agentic AI & Autonomous Systems: Multi-Agent Systems (CrewAI, LangGraph), Task Planning & Execution Agents, Tool-Calling/Function-Calling Agents, Workflow Orchestration, Reasoning Chains, Autonomous Decision Systems

•Machine Learning & Deep Learning: TensorFlow, PyTorch, Keras, Scikit-learn, XGBoost, LightGBM, Neural Networks, CNNs, RNNs, LSTMs, Attention Mechanisms, Model Optimization, Hyperparameter Tuning, AutoML

•NLP & Text Processing: Hugging Face Transformers, BERT, RoBERTa, Sentence-Transformers, spaCy, NLTK, NER, Text Classification, Sentiment Analysis, Summarization, Question Answering

•Backend & Frameworks: FastAPI, Flask, Django REST, Node.js, REST APIs, Microservices, Hugging Face Transformers

•Data Science & Analytics: Pandas, NumPy, SciPy, Statistical Modeling, Feature Engineering, A/B Testing, Predictive Analytics, Time Series Analysis & Forecasting, Revenue/Pricing Optimization, Clustering, Classification, Regression

•Cloud & MLOps: AWS (Bedrock, SageMaker, Lambda, Glue, S3, Redshift, ECS), GCP (Vertex AI, BigQuery, Dataflow, Dataproc, Document AI, Cloud Composer), Azure ML, Docker, Kubernetes, MLflow, CI/CD (GitHub Actions, Control-M), Terraform

•Data Engineering: PySpark, Hadoop, Apache Airflow, Control-M, Kafka, Pub/Sub, ETL/ELT Pipelines, Enterprise Data Warehousing (Snowflake, Databricks, BigQuery, Redshift), dbt, Data Fusion

•Visualization & Tools: Tableau, Looker, Streamlit, Power BI, Git, Jupyter, VS Code & VS Code Extensions Development, OpenAI Whisper, Speech-to-Text APIs

PROFESSIONAL EXPERIENCE

UPS — Atlanta, GA (TCS)

Gen AI Developer & ML Engineer April 2025 – Present (Remote)

Gen AI Knowledge Assistant LLM Integration RAG Systems

•Designed and implemented an Agentic AI-powered system to process, interpret, and validate enterprise policy documents across insurance, compliance, and financial domains.

•Built a Retrieval-Augmented Generation (RAG) pipeline using Lang Chain, Lang Graph, and vector databases (Pinecone/Chroma DB) for accurate policy-based question answering.

•Developed document parsing pipelines using OCR and NLP (spaCy, Hugging Face) to extract structured entities such as clauses, conditions, and policy rules.

•Implemented a multi-agent architecture (Lang Graph/Crew AI) with specialized agents for document understanding, policy validation, decision-making (approve/reject/escalate), and human-readable explanation.

•Built event-driven data pipelines (Kafka, GCP Pub/Sub) for distributed document processing and asynchronous multi-agent coordination.

•Integrated LLM-based reasoning (GPT-4/Llama) to evaluate user queries against policy rules and generate context-aware, justified decisions.

•Leveraged AWS Bedrock to host and invoke foundation models (Claude, Titan, Llama 2) for enterprise policy intelligence, enabling secure, scalable, fully managed LLM inference without managing underlying infrastructure; integrated Bedrock Knowledge Bases for RAG-based document retrieval against S3-stored policy corpora.

•Designed a hybrid rule + LLM validation layer combining deterministic business rules with LLM reasoning to improve accuracy and reduce hallucinations.

•Built an evaluation and guardrails framework including hallucination detection via grounded response checks, confidence scoring, fallback mechanisms, and prompt safety guardrails.

•Delivered a policy intelligence system integrating OCR, NLP, and LLM reasoning with hybrid rule-based and neural decision layers, achieving a 70% reduction in manual validation effort.

•Led GenAI platform scalability and resiliency testing; optimized RAG search and prompt execution, reducing average query latency by 35%.

•Deployed scalable APIs using FastAPI, containerized with Docker, and integrated with AWS/GCP for enterprise-grade production deployment.

Environment: Python, FastAPI, AsyncIO, LangGraph, CrewAI, GPT-4, AWS Bedrock (Claude, Titan, Llama 2), Kafka, ChromaDB, Pinecone, Docker, AWS (Bedrock, Lambda, ECS, S3), GCP, MLflow, Terraform

NCR Atleos — Frisco, TX AI/ML Engineer & Data Scientist Jan 2024 – March 2025 (Remote)

FinTech AI Solutions ATM Assistants LLM Integration RAG Systems

•Contributed to building a GenAI platform for FinTech, focusing on LLM-powered assistants and financial-data RAG pipelines.

•Implemented async API services and thread-safe orchestration for real-time reasoning and workflow completion via CrewAI agents.

•Architected multi-repo codebase contributions (GitHub Actions, CI/CD pipelines) with modular ML components and shared MLflow experiments; integrated Snowflake as the enterprise data warehouse for feature stores and model training datasets.

•Built Node.js-based microservices and REST APIs integrated with Python AI/ML applications for enterprise solutions and workflow orchestration.

•Designed Pub/Sub message handling via Kafka and Redis Streams, enabling low-latency data distribution across LLM microservices.

•Fine-tuned domain LLMs (GPT-4, LLaMA) using LoRA, QLoRA, and PEFT, cutting inference cost by 25% while maintaining precision.

•Supported platform automation and continuous deployment (Kubernetes + Docker) for highly available GenAI endpoints across AWS ECS/Lambda.

•Collaborated with frontend/backend teams to support API integrations using Node.js and Express.js.

•Enhanced enterprise model explainability through context-aware prompt tuning and hybrid rule + LLM decision layers; developed a VS Code extension to streamline prompt engineering workflows, enabling developers to test and iterate on LLM prompts directly within the IDE.

Environment: Python, GPT-4, Mistral, CrewAI, FastAPI, Kafka, AsyncIO, Node.js, Docker, AWS Lambda, ECS, GitHub Actions, Redis, PostgreSQL, Snowflake, VS Code Extensions

Pace University — New York, NY Research Assistant – ML/AI Focus Jul 2023 – Dec 2023

Research & Development in AI/ML NLP Model Development Technical Innovation

•Conducted research on Generative AI and LLM fine-tuning, experimenting with GPT-3.5, BERT, and RoBERTa for academic text classification and research paper summarization.

•Developed automated Python data pipelines using Apache Airflow and PySpark to ingest and process large-scale research datasets, improving analysis efficiency.

•Designed and executed NLP experiments using Vertex AI and AWS SageMaker, achieving notable performance improvements through hyperparameter optimization and ensemble techniques.

•Built reproducible ML workflows with MLflow for experiment tracking, model versioning, and automated data validation/monitoring dashboards.

•Published research findings and presented at departmental seminars on NLP applications in educational technology.

Environment: Python, TensorFlow, PyTorch, Hugging Face, BERT, GPT-3.5, Vertex AI, AWS SageMaker, PySpark, Airflow, MLflow, Pandas, NumPy, Jupyter, Git

Micron Technologies — India Data Engineer / Analyst Oct 2018 – Nov 2022

Data Pipeline & ETL Development Performance Optimization

•Built and optimized PySpark ETL pipelines processing 1TB+ of manufacturing and sensor data daily, improving data ingestion performance by 45%.

•Developed distributed data processing workflows using PySpark on AWS EMR for scalable feature engineering across millions of rows.

•Implemented deep learning models in PyTorch (CNNs and LSTMs) for defect detection and time-series forecasting, achieving a 28% improvement in anomaly detection accuracy.

•Automated Python-based data quality checks and validation frameworks, reducing manual QA effort by 60%.

•Built modular, production-ready Python utilities for feature extraction, preprocessing, and model evaluation used across multiple ML projects.

•Used PyTorch Lightning to streamline model training, early stopping, and hyperparameter tuning, cutting model iteration time by 30%.

•Integrated PyTorch models into the existing MLOps pipeline with MLflow for versioning, experiment tracking, and deployment.

•Designed Spark SQL and PySpark UDFs to handle complex business logic and real-time transformations on large-scale logs.

•Implemented parallelized hyperparameter optimization (Ray Tune + PyTorch) to train 50+ configurations across distributed AWS environments.

•Designed interactive Power BI dashboards for manufacturing KPIs, defect rates, and sensor performance metrics, enabling real-time operational insights.

•Built optimized Power BI data models using star/snowflake schemas to handle large-scale datasets sourced from BigQuery, Databricks, and Spark outputs.

•Implemented incremental refresh and query folding, reducing dashboard refresh times on multi-million-row datasets.

•Integrated Power BI with cloud data sources (GCP BigQuery, SQL databases, CSV/Parquet from data lakes) for automated, near real-time reporting.

•Designed and maintained enterprise data warehouse solutions on Snowflake and Databricks, architecting star/snowflake schemas, optimizing query performance, and delivering revenue and pricing optimization insights through advanced time series forecasting models.

•Served as primary point of contact for business stakeholders and management, presenting data-driven insights and model results; mentored junior data engineers and analysts on pipeline best practices and Python development standards.

Environment: Python, SQL, PySpark, Snowflake, Databricks, GCP (BigQuery, Dataflow, Dataproc, Vertex AI), Airflow, Control-M, Cloud Composer, XGBoost, TensorFlow, Docker, Kubernetes, GitHub Actions, Kafka, Hadoop, MLOps

KEY PROJECTS

Agentic AI Workflow System

•Designed and implemented a multi-agent AI system using LangGraph and CrewAI for autonomous task execution.

•Enabled dynamic decision-making using LLM reasoning, tool usage, and persistent memory.

•Integrated external APIs and tools to simulate real-world enterprise workflows, improving automation efficiency and reducing manual intervention.

GenAI Document Processing Pipeline

•Fine-tuned GPT-3.5 using Hugging Face and LoRA for legal document summarization and contract analysis, significantly reducing review time.

•Built end-to-end RAG pipelines using Vertex AI, LangChain, and ChromaDB for intelligent document Q&A across large repositories.

Real-Time Analytics Dashboard (AWS)

•Architected a serverless data pipeline using AWS Glue, Lambda, Kinesis, and Redshift to visualize user behavior analytics with near real-time insights.

EDUCATION

Master of Science, Data Science — Pace University, New York, NY (Jan 2023 – Dec 2024)

Bachelor of Technology — St. Mary's Institute of Technology, Andhra Pradesh, India (Mar2013 - OCT2017)

Contact this candidate