Janvitha Mandyam

Genai Ml Engineer @Citi

Denton, TX, US
MOBILE NUMBERS
+91 *********19

Signup · Get unlimited contacts

WORK HISTORY

Oct 2024 — Present

Genai Ml Engineer @Citi

View department →

Chicago, IL, US

Designed and fine-tuned prompts for GPT-3.5, GPT-4, and Claude using OpenAI APIs to automate customer service, FAQ generation, and document summarization, boosting response efficiency.• Built end-to-end Retrieval-Augmented Generation (RAG) pipelines integrating LangChain, FAISS, Weaviate, and Azure Cognitive Search, increasing enterprise Q&A accuracy by 35%.• Implemented hybrid semantic + keyword search using embeddings and LlamaIndex for document ingestion, chunking, and retrieval.• Leveraged OpenAI Function Calling and Tool Use API for multi-step reasoning and workflow automation in internal systems.• Prototyped lightweight RAG-like context chaining using BERT embeddings and NER pipelines, optimizing information retrieval prior to standardized GenAI frameworks.• Applied Guardrails.ai to enforce LLM safety, output control, and hallucination mitigation in customer-facing applications.• Developed an AI-powered migration tool converting Java Struts applications to FastAPI and React, utilizing OpenAI and Hugging Face models.• Built multi-modal pipelines for unified retrieval of PDFs, images, and structured metadata using OCR tools, Azure Form Recognizer, and LangChain agents.• Established LLMOps workflows with automated model versioning, monitoring (PromptLayer, Weights & Biases), and evaluation pipelines using EvalLangChain, TruLens, and BLEU/ROUGE metrics.• Conducted instruction fine-tuning and LoRA-based adaptation via Hugging Face, improving factual consistency and domain alignment.• Performed continuous A/B testing and evaluation of LLM variants using OpenAI Eval, TruLens, and custom scripts, optimizing prompt strategies and reducing hallucinations.• Engineered synthetic training data with GPT-4 Turbo for low-resource intents and edge-case scenarios, enhancing model generalization.• Applied prompt chaining, dynamic context window resizing, and semantic memory modules to enable multi-turn conversations and long-form query handling.

EDUCATION

N/A

University of North Texas

Master's degree, Business Analytics

2018 — 2022

Sri Venkateswara College of Engineering, Tirupati

Bachelor of Technology - BTech, Information Technology

ABOUT JANVITHA MANDYAM

I’m an AI/ML Engineer with over 4+ years of experience building and deploying machine learning solutions that make a real impact. My work focuses on data science, deep learning, NLP, computer vision, and generative AI, using tools like Python, TensorFlow, PyTorch, and Scikit-learn to turn data into intelligent, scalable systems.I enjoy solving complex problems and creating models that actually work in production. I’ve built and deployed LLM-based applications, RAG pipelines, and MLOps frameworks on AWS, Azure, and Google Cloud, ensuring every project is optimized for performance and reliability.I’m also passionate about responsible AI — using explainability tools like SHAP and LIME to make models transparent and fair. Having worked with healthcare data (Medicare, Medicaid, HL7, EDI), I understand the importance of data privacy, compliance, and accuracy.Collaboration is at the core of how I work. I love partnering with data scientists, engineers, and product teams to bring ideas to life and deliver data-driven solutions that move businesses forward. I’m always learning, exploring new advancements in LLMs, MLOps, and cloud AI, and finding better ways to connect technology with real-world value.

This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.