● Open to Work

kartikp1608

AI Engineer / Data Scientist with 1 year of experience building and deploying production AI systems across LLM applications, RAG, NLP, speech analytics, and conversation intelligence.
Gurgaon, India

About Me

AI Engineer / Data Scientist with 1 year of experience building and deploying production AI systems across LLM applications, RAG, NLP, speech analytics, and conversation intelligence. Experienced in developing scalable AI microservices, asynchronous processing pipelines, and semantic retrieval systems using Python, FastAPI, RabbitMQ, MongoDB, Vertex AI, and Docker.

Skills & Technologies

PythonJavaSQLLLMsHuggingFaceRAGLangGraphLangChainGeminiVertex AIPrompt EngineeringEmbeddingVector SearchNLPDeep LearningCNNTransformerPyTorchTensorFlowKerasSentenceTransformersStatisticsModel EvaluationFastAPIFlaskREST APIsMicroservicesEvent-Driven ArchitectureAsynchronous Backend ProcessingMongoDBMilvusMySQLSQLiteRabbitMQData Structures & AlgorithmsDBMSOperating SystemsComputer NetworksDockerGoogle Cloud PlatformGitLinuxMLflow

Work Experience

Data Scientist

ODIO
Jul 2025NOW

Developing and deploying end-to-end AI systems focusing on LLM applications, RAG, speech analytics, and scalable AI pipelines. Built and deployed solutions using Python, FastAPI, RabbitMQ, MongoDB, NLP, LLMs, Whisper, VAD, FFmpeg, Gemini, LangGraph, Milvus, SentenceTransformers, Vertex AI, Flask, PostgreSQL/pgvector, Redis, Docker, Kubernetes, and CI/CD. Designed asynchronous pipelines handling over 10,000 conversation recordings monthly and performed evaluation of LLM and RAG models to ensure quality and performance.

Projects

Skin Disease Detection (CNN)

Achieved 96% accuracy on a large dermatology dataset and reduced false negatives by 15% through targeted model improvements. Deployed a real-time inference system via Flask and Docker with an image preprocessing pipeline that cut data processing time by 40%.

PythonTensorFlowKerasFlask+3

AI Assessment Recommendation System

Built an LLM-powered recommendation engine that analyzes responses with vector similarity to produce accurate assessment suggestions in real time. Integrated front-end with HTML/CSS/JavaScript and deployed on Google Cloud for scalable, low-latency performance.

PythonPyTorchFastAPIMilvus Lite+7

Education

B.Tech

Computer Science & Engineering

Chandigarh University, Mohali2022 - 2025

Diploma

Automobile Engineering

Pimpri Chinchwad Polytechnic, Pune2016 - 2019
kartikp1608 - AI Engineer / Data Scientist with | HiringAnt