Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I’m Pravallika Bollavaram, an AI/ML Engineer with around 5 years of experience delivering data-driven solutions across retail, healthcare, and financial services. I enjoy turning complex data into practical models and insights that empower decision-makers. I’m proficient in Python, PyTorch, TensorFlow, Scikit-learn, and Hugging Face transformers, with experience building multi-modal LLM pipelines, RL agents, and time-series forecasting. I’ve enabled 90% retrieval accuracy across 80k+ internal documents and deployed production-grade ML pipelines with strong feature engineering and interpretability.…I’m Pravallika Bollavaram, an AI/ML Engineer with around 5 years of experience delivering data-driven solutions across retail, healthcare, and financial services. I enjoy turning complex data into practical models and insights that empower decision-makers. I’m proficient in Python, PyTorch, TensorFlow, Scikit-learn, and Hugging Face transformers, with experience building multi-modal LLM pipelines, RL agents, and time-series forecasting. I’ve enabled 90% retrieval accuracy across 80k+ internal documents and deployed production-grade ML pipelines with strong feature engineering and interpretability.

Pravallika Bollavaram

Data Scientist, AI Engineer, Programmer, +5





I’m Pravallika Bollavaram, an AI/ML Engineer with around 5 years of experience delivering data-driven solutions across retail, healthcare, and financial services. I enjoy turning complex data into practical models and insights that empower decision-makers. I’m proficient in Python, PyTorch, TensorFlow, Scikit-learn, and Hugging Face transformers, with experience building multi-modal LLM pipelines, RL agents, and time-series forecasting. I’ve enabled 90% retrieval accuracy across 80k+ internal documents and deployed production-grade ML pipelines with strong feature engineering and interpretability.…I’m Pravallika Bollavaram, an AI/ML Engineer with around 5 years of experience delivering data-driven solutions across retail, healthcare, and financial services. I enjoy turning complex data into practical models and insights that empower decision-makers. I’m proficient in Python, PyTorch, TensorFlow, Scikit-learn, and Hugging Face transformers, with experience building multi-modal LLM pipelines, RL agents, and time-series forecasting. I’ve enabled 90% retrieval accuracy across 80k+ internal documents and deployed production-grade ML pipelines with strong feature engineering and interpretability.

Available to hire

I’m Pravallika Bollavaram, an AI/ML Engineer with around 5 years of experience delivering data-driven solutions across retail, healthcare, and financial services. I enjoy turning complex data into practical models and insights that empower decision-makers.

I’m proficient in Python, PyTorch, TensorFlow, Scikit-learn, and Hugging Face transformers, with experience building multi-modal LLM pipelines, RL agents, and time-series forecasting. I’ve enabled 90% retrieval accuracy across 80k+ internal documents and deployed production-grade ML pipelines with strong feature engineering and interpretability.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

Work Experience

AI/ML Engineer at JP Morgan Chase &Co.

February 1, 2025 - Present

Developed generative AI models using OpenAI GPT-4 and Hugging Face Transformers for internal market research summarization, producing 500 structured reports/month and reducing analyst prep time by 35%. Built portfolio risk prediction models using PyTorch-based graph neural networks, analyzing 9,000+ client portfolios with 42+ risk features, enabling daily risk scoring across 5 asset classes. Implemented reinforcement learning agents with Ray RLlib to optimize intraday trading strategies, simulating 1,200 trade sequences/day, evaluating 6 performance metrics, and improving backtested PnL by 7%. Engineered real-time data streaming pipelines using Kafka and Spark Structured Streaming, processing trade events with low latency and high throughput while maintaining audit compliance. Performed advanced feature engineering and model interpretability analysis using SHAP, ELI5, and Optuna, generating 25+ explainability charts and tuning 10 model configurations.

Machine Learning Scientist at CVS Health

May 1, 2024 - January 31, 2025

Designed and implemented a Retrieval-Augmented Generation (RAG) pipeline using LLMs, embeddings, and vector databases to enable secure Q&A over 120k internal healthcare documents. Programmed GenAI orchestration workflows with LangChain to support care coordination, policy lookup, and member support scenarios. Constructed data ingestion and transformation pipelines with Apache Spark, improving data availability for downstream ML workflows and cutting end-to-end processing latency by 28%. Trained and evaluated predictive models using XGBoost on claims and eligibility data for 12–15 months of history. Established experiment tracking and model governance with MLflow across 30+ training runs to ensure reproducibility and audit readiness. Collaborated with product, clinical, and engineering teams in a HIPAA-compliant environment, delivering 6+ production ML features from design to deployment.

Data Scientist at Accenture

January 1, 2020 - December 31, 2022

Architected demand forecasting models using Python, Pandas, NumPy, and classical time-series techniques for 3 product lines across 45 stores, enabling weekly SKU-level planning. Built supervised ML models with Scikit-learn (KNN) for 18 months of order and inventory data to identify slow-moving items and inform replenishment decisions. Applied customer purchase propensity modeling using Deep Learning (TensorFlow, LSTM) to capture sequential buying behavior for promotion-driven campaigns. Performed data cleaning, validation, and feature engineering on POS and warehouse datasets to deliver analysis-ready data. Orchestrated NLP on customer feedback and product descriptions to extract demand signals for merchandising and assortment analysis. Supported model deployment and monitoring by packaging models for internal analytics platforms and validating inference outputs during retraining cycles.

AI/ML Engineer at JP Morgan Chase & Co.

February 1, 2025 - Present

Developed generative AI models using OpenAI GPT-4 and Hugging Face Transformers for internal market research summarization, producing 500 structured reports per month and reducing analyst prep time by 35%. Built portfolio risk prediction models using PyTorch-based graph neural networks across 9,000+ client portfolios with 42+ risk features, enabling daily risk scoring across 5 asset classes. Implemented reinforcement learning agents with Ray RLlib to optimize intraday trading strategies, simulating 1,200 trade sequences per day and improving backtested PnL by 7%. Engineered real-time data streaming pipelines using Kafka and Spark Structured Streaming with low latency and audit-compliant monitoring. Conducted feature engineering and model interpretability analysis with SHAP, ELI5, and Optuna, generating 25+ explainability charts and tuning 10 configurations.