
Harsh Patel
AI ML Engineer for RAG Chatbots, LLM Agents and Computer Vision
Habilidades

Revisa mis servicios


Porfolio
Experiencia laboral
Confidential
Tiempo completo • 2 yrs 7 mos
AI ML Engineer
Aug 2024 - Present • 2 yrs 2 mos
AI/ML Engineer building production generative AI and computer vision systems at a software company (name withheld under NDA). Promoted from intern in Aug 2024. Selected work: - Conversational AI Agent Platform: LLM agents for WhatsApp and web chat with personas, skills, guardrails and tool calls to CRM, calendar and RAG/SQL knowledge bases. Built its evaluation layer (rule-based defect detectors + LLM-as-judge) that triages live conversations and drives prompt improvements. - RAG Virtual Assistant: document Q&A with source citations and guardrails (LangChain, OpenAI, vector search, FastAPI). - Document & Image AI: OCR and YOLO detection/segmentation pipelines for scanned documents and images (PaddleOCR, Google Cloud Vision, Ultralytics). - Generative AI Chat App: streaming chat plus image generation through OpenRouter and Replicate. - Web Data Extraction: Selenium scraping pipelines that turn websites into structured data. Stack: Python, FastAPI, Flask, LangChain, OpenAI, Gemini, FAISS, ChromaDB, Elasticsearch, Redis/RQ, Docker, AWS S3, Firebase, Google Cloud, PostgreSQL.
AI ML Engineer Intern
Feb 2024 - Jul 2024 • 5 mos
AI/ML internship at a software company (name withheld under NDA), promoted to full-time AI/ML Engineer in Aug 2024. - Trained and evaluated YOLO object detection and segmentation models on custom datasets labelled with CVAT and Label Studio. - Built OCR experiments for scanned documents and images (PaddleOCR, Google Cloud Vision). - Prototyped LLM and NLP features with OpenAI and LangChain, and wrapped them in Python FastAPI and Flask APIs. - Wrote Selenium scrapers and data-cleaning scripts with Pandas and NumPy. Stack: Python, PyTorch, Ultralytics YOLO, OpenCV, PaddleOCR, OpenAI, LangChain, FastAPI, Flask, Docker, PostgreSQL.