Open to opportunities

Swaraj Mali

@swarajmali1

Message

I build AI-powered systems in NLP, RAG, and computer vision—turning documents and data into actionable insights.

India

Message

What I'm looking for

I’m looking for an entry-level AI/ML role where I can ship RAG and computer-vision systems end-to-end, work with strong engineers, and keep learning through real product ownership.

I’m a B.Tech Computer Science student who loves building end-to-end AI products—especially where NLP, retrieval (RAG), and computer vision meet real-world workflows. In Deloitte’s Capstone 2025, I helped build an AI-powered audit platform and we were shortlisted in the top 30 out of 200+ teams.

I apply modern LLM pipelines to automate messy information: I leveraged LLaMA 3 with RAG, OCR, ChromaDB, and Redis to analyze financial documents, flag inconsistencies, and generate audit-compliant reports in real time. I’m equally comfortable bridging AI with practical engineering—integrating APIs and designing robust pipelines that actually work.

Recently, in my AI/3D internship experience, I built an interactive AI avatar system with Azure Cognitive Services for real-time Speech-to-Text, Text-to-Speech, and viseme-driven lip-syncing. I also contributed computer vision work—training deep learning models (FCN, DeepLabV3), using YOLO, and creating a similarity-search pipeline with custom feature extractors (Gabor, FFT).

Experience

Work history, roles, and key accomplishments

AI/3D Intern

Streamoji

Oct 2025 - Dec 2025 (2 months)

Integrated Azure Cognitive Services to build an interactive AI avatar with real-time Speech-to-Text, Text-to-Speech, and viseme-driven lip-syncing. Trained deep learning models for face semantic segmentation and built a computer-vision pipeline (YOLO + feature extractors) to power vector-based similarity search for matching facial features and hairstyles.

Azure Cognitive Search Text To Speech Viseme Lip Sync Face Semantic Segmentation YOLO Similarity Search Computer Vision DeepLabv3 FFT

Capstone Program Participant

Deloitte

Apr 2025 - Jul 2025 (3 months)

Built an end-to-end AI-powered audit platform for Deloitte Capstone 2025, shortlisted in the top 30 out of 200+ teams. Used LLMs with RAG plus OCR and vector search to automate financial document analysis, flag inconsistencies, and generate audit-compliant reports in real time.

RAG ChromaDB Redis Vector Search Document Analysis Audit Reporting