Haveela Karre
Verified Expert in Engineering
Data Scientist and ML Developer
Vijayawada, Andhra Pradesh, India
Toptal member since October 21, 2025
Haveela is a data scientist with three years of experience developing AI-driven solutions that transform data into actionable insights. She specializes in machine learning (ML), large language models (LLM), and predictive analytics and has made impactful contributions at Mavvrik.ai, Turing, and Tredence Inc. Haveela's work drives data innovation and enhances business intelligence.
Portfolio
Experience
- Python 3 - 5 years
- PyTorch - 5 years
- SQL - 5 years
- Machine Learning - 4 years
- Retrieval-augmented Generation (RAG) - 3 years
- Large Language Models (LLMs) - 3 years
- Google BigQuery - 3 years
- LangChain - 3 years
Preferred Environment
Windows, Python 3, Slack
The most amazing...
...things I've developed are agentic text-to-SQL frameworks, LLM pipelines, and predictive analytics models that improved efficiency by 80%.
Work Experience
AI Engineer – RAG & Agentic AI
PepsiCo Global - MD Projects
- Developed and scaled PepGPT, PepsiCo’s enterprise-grade AI assistant platform powering conversational analytics, intelligent enterprise search, multi-LLM orchestration, and secure business-data interactions across connected enterprise systems.
- Integrated Azure AI Foundry, Microsoft Fabric, Power BI Semantic Models, enterprise connectors, and agent-based workflows to enable production-ready AI experiences for internal users.
- Led back-end orchestration, authentication flows, runtime agent execution, connector onboarding, and UI enhancements while fixing critical production issues and delivering new AI-driven functionalities across the platform.
- Developed the PepGPT platform by implementing a Skills feature inspired by Claude Skills, and enhanced the platform by integrating Azure DevOps Boards to enable seamless project and work item management.
AI Engineer
Turing
- Developed and fine-tuned large language models (LLMs) using supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF), improving instruction adherence and contextual accuracy by 25%.
- Implemented retrieval-augmented generation (RAG) pipelines integrating LlamaIndex and LangChain, enabling accurate, context-aware responses for enterprise-scale knowledge systems.
- Optimized model inference performance by reducing latency by 30% through prompt optimization, efficient context management, and caching strategies.
Data Science Consultant
Mavvrik AI
- Developed an agentic AI text-to-SQL framework using VannaAI and BigQuery, automating over 80% of manual reporting tasks.
- Transformed Datastore data into structured SQL outputs, delivering dynamic Excel-ready insights for client decision-making.
- Strengthened cloud-native data intelligence pipelines using BigQuery, Cloud Storage, and Python to improve automation and transparency.
- Enhanced cross-team accessibility through AI-driven reporting templates that reduced turnaround time by 60%.
Data Scientist
Tredence
- Developed predictive detention-risk models with XGBoost, Random Forest, and CatBoost on Databricks that cut logistics detention costs by around 15%.
- Built end-to-end ETL pipelines in PySpark and Azure Data Lake to automate data ingestion and model retraining for real-time monitoring.
- Performed advanced EDA and root-cause analysis on large supply-chain datasets to uncover key drivers of out-of-stock and dwell-time issues.
- Created interactive Power BI dashboards enabling stakeholders to visualize model outputs and act on predictive insights instantly.
Experience
Custom Reports | BigQuery GCP
https://www.mavvrik.ai/Text-to-SQL Using Agentic AI
https://www.mavvrik.ai/Predictive Detention Risk Model
AI Engineer Work
Education
Bachelor's Degree in Data and Computer Science
KL University - Guntur, India
Skills
Libraries/APIs
PyTorch, PySpark, OpenAI Assistants API, Python Asyncio, REST APIs, OpenAI API, Claude API, TensorFlow, Pydantic
Tools
Slack, BigQuery, Azure OpenAI Service, AWS Cloud Development Kit (CDK), Claude Code, ChatGPT, Claude Agent SDK, Celery, Amazon Elastic Container Service (ECS), Amazon SageMaker, Amazon CloudWatch, Claude, Codex
Languages
Python 3, SQL, Python, Snowflake, TypeScript, Excel VBA
Frameworks
Flask, LangGraph, Django, Streamlit, Agentic Frameworks, LlamaIndex
Paradigms
ETL, Microservices, Testing, Business Intelligence (BI), Real-time Systems, Model Context Protocol (MCP), Azure DevOps
Platforms
Amazon Web Services (AWS), Docker, Jupyter Notebook, Microsoft Fabric, Databricks, Azure, AWS Lambda, Windows
Storage
NoSQL, JSON, Amazon S3 (AWS S3), PostgreSQL, Graph Databases, Databases
Industry Expertise
Project Management
Other
Large Language Models (LLMs), Retrieval-augmented Generation (RAG), Machine Learning, Google BigQuery, LangChain, Artificial Intelligence (AI), Generative Artificial Intelligence (GenAI), AI Modeling, AI Model Integration, APIs, Agentic AI, AI Agents, FastAPI, Pinecone, Data Analysis, Data Analytics, Data Science, Software Development Lifecycle (SDLC), Vector Search, Natural Language Processing (NLP), Predictive Analytics, Cloud Services, Document Processing, AI Chatbots, Multimodal GenAI, Containerization, Architecture, Product Development, Prompt Engineering, Recommendation Systems, Amazon Bedrock AgentCore, RAG Pipelines, OpenAI, Design, Reliability, Dashboards, Data Visualization, CI/CD Pipelines, Deep Learning, Open-source LLMs, API Integration, Data Modeling, Computer Vision, Payment APIs, Payments, ETL Pipelines, AI Copilots, ChatGPT API, ChatGPT Prompts, IT Projects, IT Project Management, Agentic RAG Systems, Vector Databases, Solution Architecture, Data Privacy, Identity & Access Management (IAM), Neural Networks, Fine-tuning, Business Analysis, AI Model Training, Machine Learning Operations (MLOps), Platform as a Service (PaaS), Azure Databricks, PDF Scraping, Serverless, Model Tuning, Data Architecture, Data Engineering, Large Language Model Operations (LLMOps), Chatbots, Software Engineering, Personally Identifiable Information (PII), Cloud Platforms, Machine Learning (ML) APIs
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring