Anvesha Mittal

Anvesha Mittal

AI/ML Intern

Artificial Intelligence and Machine Learning
2roles
38skills
2education

About

B.Tech Computer Science student at Birla Institute of Technology, Mesra, with an 8.59 CGPA and internship experience building production-ready RAG systems, semantic search pipelines, LLM formatting systems, and high-throughput data processing engines.

Experience

AI/ML Intern

Shoro AI Labs

Dec 2025 - Feb 2026

Engineered a two-phase LLM formatting pipeline that transformed streamed AI output into deterministic Markdown for 100% of user queries. Implemented multi-pass Markdown normalization, comparison-question JSON-to-table rendering using SSE, automatic subject routing with LLM intent detection and vector similarity scoring, and contributed through Git branches and reviewed pull requests.

AI/ML Intern

Bootcoding Pvt. Ltd.

Nov 2025 - Dec 2025

Generated structured interview transcripts in JSON/JSONL schemas, cleaned and normalized job-description data, built an API-based document ingestion pipeline, implemented Qdrant semantic search with embeddings and chunking, and engineered processing for 150–200 concurrent inputs and 7,600+ documents with schema validation and error recovery.

Education

Birla Institute of Technology, Mesra

Bachelor in Computer Science Engineering, Computer Science Engineering

2023 - 2027

CGPA: 8.59

Bhartiya Vidya Bhavan Vidyashram, Jaipur

Class 10th and Class 12th

2012 - 2023

Class 10th: 91%; Class 12th: 84.6%

Skills

C++CPythonData Structures & AlgorithmsObject-Oriented Programming (OOP)Operating SystemsDBMSRAGVector DatabasesEmbeddingsRegressionClassificationClusteringNeural NetworksFastAPILangChainTensorFlowScikit-learnDockerGitAzure OpenAIFAISSGit WorkflowsPull RequestsCode ReviewsLangGraphQdrantSSEJSONJSONLAzure BlobPyTorchPandasNumPyTF-IDFRidgeGridSearchCVSMAPE

Projects

Opportunity Radar AI

Built an AI-native stock analysis platform for Indian retail investors using a five-agent LangGraph pipeline with Researcher, Bull, Bear, Judge, and Alert agents. Engineered a Divergence Detector comparing FII/DII institutional flows with retail social sentiment to flag high-risk trap signals.

RAG System

Built a production RAG chatbot with FastAPI and Azure, supporting real-time knowledge-base updates through asynchronous ingestion. Reduced query latency by 30–40% through FAISS index tuning and async I/O.

Smart Product Price Prediction

Built an end-to-end ML pipeline for more than 75,000 listings using regex-based feature extraction. Trained a TF-IDF and Ridge model with GridSearchCV and SMAPE, achieving a 40.16 validation score.