Taaniya Arora

WORK HISTORY

EDUCATION

ABOUT TAANIYA ARORA

Focused on research and development and building state of the art Agentic AI products and platforms with 8.5+ yrs experience in AI/ML and data science.Self-driven, creative & passionate about data science, machine learning, deep learning, natural language processing, reinforcement learning, problem solving, research and learning advancements in the field.Currently growing in each of these areas and gaining expertise in solving complex problems for building smart and intelligent products & their features. I enjoy doing research and solving challenging problems with ML and AI techniques with a holistic approach.First author AI/ML research paper publications in proceedings of conference -https://aircconline.com/csit/papers/vol11/csit112313.pdfhttps://adasci.org/lattice-volume-4-issue-1/intelliqsense-an-intelligent-real-time-query-autocompletion-framework-using-gpt-2/GenAI based-features I\'ve contributed to - Building multi-agentic systems- RAG (Retrieval + Reranking + Generation)- Intent classification, orchestration and planning for multi-agent collaboration- LLM-as-a-judge / Critic- Query understanding & query rewriting- Conversation management in conversational AI interface- Schema linking, Named Entity Recognition and handling ambiguities in Enterprise Text-to-SQL system- Fine-tuning and evaluation of SBERT model for semantic retrieval on task-specific custom dataset with benchmarking different embedding models on eval metrics - MRR, precision@k, recall@k- Built a novel fine-tuned GTP2 model-based on task specific custom dataaset for autocompletion feature for conversation AI interface- Built a custom evaluation framework for natural language generated output from fine-tuned GPT2 for autocompletion. Tools & TechnologiesAzure OpenAI, Groq AI service, Azure AI search, LangSmith, OpenAI, Docker, Kubernetes, Azure DevOpsLLMs - Open AI GPT2, GPT 3.5, GPT 4 / Turbo / 4o, Llama 3, deepseek r1 Libraries -ML, Deep learning, Data Science & Research-Huggingface transformers, pytorch, keras, tensorflow, scikitlearn, OpenCV, numpy, pandas, matplotlib, seaborn, bokeh, plotlyDistributed big data processing & ML-SparkML, pySpark, Libraries for Building Multi-agent systems, GenAI apps, RAGs -LangGraph, AutoGen 0.5.x, 0.2, Langchain, LlamaIndex, CrewAI, TaskWeaver, sentence-transformers, ranxDatabases -FAISS, Weaviate, PostGreSQL, MySQL, HDFS, Hive Warehouse.Version control -GitHub, GitLab, AzureDevOps