×
About Skills Experience Projects Certs Contact
Open to AI Engineering roles · Thane, IN

Hi, I'm Ujwala.

I build and ship production-grade AI systems — from multi-agent LLM pipelines and computer-vision inspection to scalable, GPU-accelerated deployment. I turn research into products that deliver measurable business impact.

View My Work Get in Touch
0
AI Roles
0
Shipped Projects
0
Certifications
0
Tools & Frameworks
SCROLL
01 · About

Engineering intelligence into products

I'm an AI Engineer currently leading research and solution architecture for an AI-powered vehicle damage inspection system at USA Direct Auto Drivesoft, combining Computer Vision, LLMs, and OpenAI Vision into a production roadmap.

Previously, as an AI Engineer Intern at Sminetech Solutions, I built and deployed AI platforms for market intelligence, document verification, and outreach automation — designing end-to-end pipelines for unstructured multimedia and document data.

I care about the whole lifecycle: data annotation, model architecture, training pipelines, deployment, and the dashboards that make it all useful to the business.

// profile

Focus
LLMs, Agents & Computer Vision
Location
Thane, India
Education
MCA · SVIMS, Pune
Currently
AI Specialist @ Drivesoft
02 · Toolkit

Skills & Technologies

03 · Journey

Experience

Feb 2026 — Jul 2026

AI Specialist

USA Direct Auto Drivesoft
  • Led research and solution architecture for an AI-powered vehicle damage inspection system using Computer Vision, LLMs, and OpenAI Vision.
  • Owned the technical roadmap — data annotation strategy, model architecture, training pipelines, AWS infrastructure, and deployment planning.
  • Evaluated & benchmarked YOLO11, Roboflow, FastAPI, Docker, AWS, and OpenAI Vision for scalable, production-ready solutions.
  • Defined AI workflows for part detection, damage detection, image classification, and structured damage reporting.
Computer VisionYOLO11OpenAI VisionAWSDocker
Jul 2025 — Jan 2026

AI Engineer Intern

Sminetech Solutions
  • Built and deployed AI-driven platforms for market intelligence, document verification, and outreach automation using FastAPI, Streamlit, Redis, LLMs, Docker, Celery, and Linux.
  • Designed end-to-end data pipelines for unstructured multimedia & document data — enabling sentiment analysis, competitor gap detection, semantic validation, and influencer scoring.
  • Developed scalable dashboards and production automation workflows with GPU acceleration and asynchronous processing.
FastAPILLMsCeleryRedisStreamlit
04 · Selected work

Featured Projects

Project 01

AI Agents for Medical Report Analysis

A multi-agent system using LangChain and Gemini to analyze medical reports (PDF/TXT) with domain-specific agents. Includes a Streamlit dashboard for real-time analysis, visualization, and report downloads.

LangChainGeminiStreamlitMulti-Agent
Project 02

AI-Powered Fabric Matcher

An image similarity system built on MobileNetV2 embeddings and cosine similarity, served through a Flask REST API with a lightweight SQLite vector store for fast retrieval.

MobileNetV2FlaskSQLiteVector Search
Project 03

Mood-Based Movie Recommender

A voice-enabled recommendation platform leveraging Whisper for speech recognition and an emotion-classification model to generate personalized movie suggestions from a user's mood.

WhisperNLPEmotion AIStreamlit
05 · Credentials

Certifications

HF
Fundamentals of Agents
Hugging Face
AC
Data Analytics & Visualization
Accenture Job Simulation
K
Intro to Machine Learning
Kaggle
K
Intermediate Machine Learning
Kaggle
K
Feature Engineering
Kaggle
K
Advanced SQL
Kaggle
06 · Contact

Let's build something intelligent.

I'm open to AI Engineering opportunities and collaborations. Whether it's LLM agents, computer vision, or shipping ML to production — let's talk.