Headline

AI/ML Engineer @ Inspiron Labs

About

I am a data scientist with experience in data analysis, machine learning, big data, and AI, including generative AI. I work with Python, SQL, and modern data tools to develop solutions that solve real-world problems and support business goals. My background includes building AI models, developing and optimizing APIs, and creating scalable systems. I focus on practical implementation, from data pipelines to model deployment, and work well in collaborative environments to deliver end-to-end solutions.

Location

Bengaluru

AI stack

claude code

langchain

langgraph

langsmith

livekit

DSPy

Work experience

Artificial Intelligence Engineer at InspironLabs, Feb 2025 - Present · 1 y 8 mo, Bengaluru, Building production-grade agentic AI systems, multimodal applications, and scalable data platforms across healthcare, finance, and commute domains. Enterprise AI Data Fabric Platform | LangGraph, DSPy, FastAPI, Postgres, Elasticsearch • LangGraph orchestration combining LLM-generated tasks and predefined workflows, with dependency-aware parallel execution across SQL, RAG, and API sources. • Document retrieval pipeline using Elasticsearch vector search, semantic chunking, citation tracking, and Redis caching for long-form PDFs. • Multi-turn clarification via DSPy signatures and LangGraph interrupts to catch incomplete requests before execution. • SSE-based token streaming for real-time incremental LLM responses, reducing perceived latency. AI Voice Agent for Healthcare | LiveKit, GraphQL, Postgres • Voice-and-chat agent for appointment booking, cancellation, rescheduling, and info retrieval. • LiveKit real-time audio streaming integrated with LLM services for voice and text. • Agent tools connected to backend appointment operations via GraphQL APIs. • Human handover between AI and admin, with intent detection, multi-step workflows, and timezone handling. AI-Powered Vision & Video Generation Platform | FastAPI, Embeddings, FFmpeg, Remotion Lambda, PostgreSQL • Conversational workflow gathering user preferences to generate personalized narratives and video scripts. • Semantic media retrieval using OpenAI CLIP embeddings to match stock assets to narrative sections via similarity search. • LAION CLAP embeddings for text-audio similarity matching to rank music against narrative mood. • JSON-based scene timelines rendered via Remotion Lambda for parallelized cloud rendering, cutting render time. Healthcare Data ETL Pipeline | Apache Airflow, AWS Glue, Snowflake, DOMO • Maintained Airflow/AWS Glue pipelines for healthcare data ingestion and transformation, and supported DOMO dashboard reporting on Snowflake.

Education

Modern Data Science and ML with specialisation in GenAI, Online, Scaler

Bachelor's degree, Computer Science, amity University