Gen AI Engineer with Production-Grade Systems Expertise

Akhila Boddu

Gen AI Engineer

**********@*****.*** +1-502-***-**** https://www.linkedin.com/in/akhila-boddu-392820275/

Professional Summary:

GenAI Developer with almost 5 years of experience designing and deploying production-grade AI systems across healthcare, finance, and enterprise data environments.

Specialized in building LLM-driven applications including RAG chatbots, multi-agent systems, and AI assistants that integrate structured retrieval, accurate grounding, and consistent reasoning.

Strong expertise implementing LLM safety, moderation, and governance using guardrails, rule engines, and vector-based classifiers to ensure secure and policy-aligned AI behavior.

Proven ability to create high-quality embedding pipelines using OpenAI, HuggingFace, SentenceTransformers, FAISS, and Pinecone, enabling reliable semantic search and contextual retrieval.

Experienced in architecting end-to-end RAG ecosystems involving chunking strategies, ranking models, retrieval consistency checks, and verification layers to improve factual accuracy.

Skilled in LLM finetuning and optimization using LoRA, QLoRA, RLHF, PEFT, and PyTorch, with hands-on work across GPT-4, LLaMA, T5, and other transformer-based models.

Background building full-stack AI applications, using FastAPI for scalable inference APIs and React for real-time conversational interfaces.

Strong ML engineering foundation from prior experience at Empower, with a track record in training models using TensorFlow, PyTorch, scikit-learn, ONNX, and end-to-end MLOps practices.

Experienced implementing automated ML workflows with Airflow, MLflow experiment tracking, SageMaker Pipelines, continuous training, and full lifecycle model governance.

Deep data engineering expertise from Samsung, including building AWS cloud data platforms, ETL/ELT pipelines with Glue, Kinesis, Lambda, EMR, Spark, and ensuring scalable data processing for analytics and ML.

Proven ability to containerize and deploy workloads using Docker and operate production systems on Kubernetes, ensuring reliability, autoscaling, and zero-downtime updates.

Adept at collaborating across engineering, data science, and security teams to deliver robust, secure, and business-aligned AI solutions that meet enterprise performance, accuracy, and compliance expectations.

Technical Skills:

GenAI & LLM Ecosystem: GPT-4, GPT-3.5, LLaMA, T5, Falcon, Mistral, Hugging Face Transformers, SentenceTransformers, LangChain, LangChain Agents, CrewAI, Guardrails AI, Prompt Engineering, RAG Architectures, Multi-Agent Systems, RLHF, LoRA, QLoRA, PEFT.

Embeddings & Vector Retrieval: OpenAI Embeddings, HuggingFace Embeddings, Pinecone, FAISS, ChromaDB, ColBERT, DPR, Semantic Search, Ranking Models.

NLP & Deep Learning: SpaCy, TensorFlow, PyTorch, ONNX Runtime, OCR Models, Transformer Models, Text Summarization, Entity Extraction, Feature Engineering.

Machine Learning & Data Science: scikit-learn, XGBoost, LightGBM, Pandas, NumPy, SciPy, TensorBoard, Model Evaluation, Feature Selection, Baseline Modeling.

MLOps, Automation & Experiment Tracking: MLflow, DVC, Airflow, SageMaker Pipelines, SageMaker Training/Deployment, Model Versioning, Automated ETL/ELT Pipelines.

Backend, APIs & Full-Stack Development: FastAPI, Flask, Node.js, React, REST APIs, Real-time Chat Interfaces, Microservices.

DevOps, Deployment & Infrastructure: Docker, Kubernetes, GitHub Actions, Jenkins, CI/CD Pipelines, Infrastructure-as-Code, Container Orchestration.

Cloud, Data Engineering & Databases: AWS (EC2, S3, VPC, Glue, Lambda, Kinesis, EMR, IAM, CloudWatch), Spark, PySpark, Redshift, Snowflake, SQL Server, DynamoDB, Cassandra, Oracle, Hive.

Education:

Master of Science (M.S.) in Engineering Management

Indiana Tech University Fort Wayne, IN USA Apr 2024

Bachelor of Technology (B.Tech.) in Electronics and Communications Engineering

Malineni Lakshmaiah Women's Engineering College, Guntur, India

Certifications:

Azure Data Engineer Master Class – Udemy

SQL Bootcamp

The complete GEN AI engineering course - Udemy

Professional Experience:

Elevance Health Oct 2024 – Current

GEN AI Developer Indianapolis, IN

Implemented enterprise-grade LLM safety controls by combining AI, a custom policy rule engine, and vector-based filtering, ensuring compliant, secure, and toxicity-reduced model outputs for production environments.

Built advanced embedding pipelines using OpenAI Embeddings for general semantic reasoning, HuggingFace models for domain-tuned representations, and Sentence Transformers for efficient similarity scoring, enabling high-accuracy semantic search and document retrieval.

Developed robust RAG ecosystems by orchestrating retrieval and generation workflows with LangChain, applying vector indexing through Pinecone for real-time search, and using FAISS for scalable offline or on-premise retrieval to support domain-heavy documentation systems.

Executed model finetuning and optimization using LoRA, QLoRA, and RLHF techniques across models such as GPT-4, LLaMA, and T5, improving accuracy while reducing inference cost and hardware requirements.

Enhanced retrieval quality by integrating ranking models like ColBERT and DPR, establishing retrieval consistency checks that strengthened grounding, improved factual accuracy, and reduced hallucination rates in RAG pipelines.

Strengthened NLP preprocessing and knowledge-extraction workflows with SpaCy, applying entity extraction, text classification, and summarization techniques to improve upstream context quality before LLM generation.

Implemented multi-agent AI systems using LangChain Agents and CrewAI, leveraging coordinated orchestration logic to automate multi-step workflows, cooperative task handling, and autonomous decision-making across complex pipeline stages.

Delivered full-stack AI assistant applications built on FastAPI for low-latency backend inference and React for interactive, real-time conversational UI, enabling end-to-end deployment of scalable, user-facing chatbot solutions.

Containerized and deployed AI services using Docker to maintain consistent runtime environments and Kubernetes to achieve scalable, resilient orchestration of LLM and RAG workloads across distributed infrastructure.

Automated the entire GenAI deployment lifecycle using CI/CD pipelines, ensuring continuous testing, validation, versioning, and rollout of prompts, embeddings, retrieval logic, and model updates with minimal manual intervention.

Created internal AI productivity tools leveraging MLflow for experiment tracking, DVC for dataset versioning, and Python-based evaluators for prompt debugging, retrieval testing, and model comparison, accelerating data team workflows and experimentation velocity.

Conducted deep root-cause analysis of LLM misbehavior, embedding drift, retrieval failures, and grounding issues using debugging utilities from LangChain, vector-store inspection tools, and embedding visualization methods to systematically enhance system reliability.

Empower Feb 2023 – Sept 2024

AI/ML Engineer Boston, MA

Fine-tuned LLMs using Hugging Face Transformers, PyTorch, and PEFT/LoRA, applying these frameworks to adapt large models to new tasks, reduce training cost, and deliver more accurate summarization and reasoning outputs.

Built reliable GenAI applications through careful prompt engineering, where I rewrote instructions to shape model behavior, and used guardrail frameworks to enforce policies that prevented hallucinations and unsafe responses.

Developed retrieval-augmented pipelines by orchestrating document flows with LangChain and indexing embeddings through FAISS/Pinecone, ensuring LLMs always referenced the correct context and produced grounded answers.

Trained and evaluated machine learning models with scikit-learn, leveraging it for baselines and quick experimentation to find the best-performing algorithms before scaling up to deep learning solutions.

Created advanced NLP and extraction workflows using TensorFlow and OCR/ONNX models, applying them to pull structured information from unstructured text and images while improving accuracy through custom preprocessing.

Managed the full ML lifecycle using MLflow to track experiments and SageMaker Pipelines to automate training, validation, and deployment, making each model version reproducible and easy to roll back or promote.

Automated training and data workflows using Airflow, defining scheduled tasks that reliably prepared data, retrained models, and pushed new versions into staging or production without manual intervention.

Cleaned, joined, and engineered large datasets using Pandas, writing robust preprocessing logic to ensure data consistency, reduce noise, and maintain high-quality training inputs across multiple pipelines.

Processed high-volume datasets with Spark, distributing transformations and feature engineering steps across clusters to handle workloads that were too large or slow for single-machine workflows.

Packaged AI models as self-contained services using Docker, creating predictable runtime environments that made deployment seamless across development, testing, and production systems.

Operated scalable inference systems on Kubernetes, using its autoscaling, load balancing, and rolling deployment capabilities to keep AI services stable, responsive, and highly available under varying workloads.

Samsung India Jun 2020 – July 2022

Associate Data Engineer Hyderabad, IN

Built secure cloud environments on AWS by configuring VPCs, isolating workloads with public and private subnets, and managing routing through route tables, NAT Gateways, and Internet Gateways. Applied Security Groups for controlled access and used Elastic Load Balancers to distribute workloads reliably.

Designed and scaled data systems using EC2 for compute, Auto Scaling to handle variable workloads, and S3 as the primary data lake. Implemented strong security boundaries with IAM and automated small data processes using Lambda functions.

Developed automated ETL pipelines with AWS Glue, building transformation logic inside Glue Jobs and managing schema consistency through the Glue Data Catalog. Set up workflow triggers to coordinate multi-step ETL/ELT operations.

Built streaming and ingestion workflows on AWS using Kinesis for event capture and Lambda for near-real-time validation and transformation before storing data in S3 for analytics and downstream processing.

Processed large-scale datasets using Apache Spark and PySpark on EMR, optimizing shuffles, partitions, and caching behavior to ensure efficient distributed computation and cost-effective cluster usage.

Implemented continuous data processing with Spark Streaming, enriching and validating data as it arrived and using checkpointing and structured-streaming principles to maintain reliability and recoverability.

Strengthened data quality by writing validation routines in Python and PySpark, performing schema checks with Hive, and applying frameworks like Great Expectations to detect anomalies and enforce data rules before downstream use.

Monitored and troubleshot pipelines through CloudWatch, reviewing Glue and EMR logs, tuning configurations, and setting up alerts so operational issues could be identified and resolved quickly.

Managed and optimized data warehouses such as Redshift, Snowflake, and SQL Server, improving performance through indexing strategies, optimized table designs, and efficient partitioning or distribution configurations.

Worked with storage and database systems like DynamoDB, Oracle, and Cassandra, adjusting data models and query patterns to support high-volume operations and fast retrieval.

Maintained reliable deployments using CI/CD practices, incorporating infrastructure-as-code, automated build pipelines, and version control to ensure consistent delivery of data jobs and cloud resources across environments.

Contact this candidate