Cere Labs

Solutions Specialist

1. Queryfic (Python, MySQL, FastAPI, OpenAI, LLM, Qdrant database, Docker)
   1. Framework for document reading, text extraction, preprocessing, and generating small segments of documents.
   2. Generates and stores vectors in a Qdrant (vector) database, generates prompts, and retrieves results for the user query.
   3. Managed a team of 8, assigned tasks, and monitored project roadmap.

2. Developed connectors and Error report service for Toyota Financial Service of India (TFSI) (Python, MySQL, FastAPI, RabbitMQ)
   - Currently processing 8,000+ pages of document on a daily basis.
   1. Folder Listener - Retrieves files from nested folders available at a mount point, maintain visited folders/files, send files for processing, and overrides original files with processed ones.
   2. Omnidocs (DMS): Retrieves files from Omnidocs using RestAPIs, build file path recursively, stores metadata, sends files for processing, and overrides originals.
   3. Error Report: Generates daily reports (Success/Failure) on assigned schedule, sends as an Excel file to designated email.

3. Bank Statement Reader (Python, MySQL, Machine Learning, Pandas)
   1. Enhanced the reading mechanism of bot by adding the capability of reading different bank formats.
   2. Improved the processing time by ~60% for processing the bank statement.
   3. Implemented machine learning model for table and column detection.
   4. Further implemented an algorithm for forming transactions spread across multiple lines.

4. NACH Mandate Reader (Python, MySQL, Azure Services)
   1. Built a bot to read handwritten NACH mandate forms using Azure OCR handwriting service.
      2. Performed object detection to identify key fields, and display the output as key-value pairs in a Excel sheet.

5. Dtect Multi Input API (Python, FastAPI, MySQL)
   1. Improved the current API, by adding the capability of taking multiple documents as input.
   2. Processing the documents by sending them to respective bots.

Software Developer

DBIT SevaTech

Developed a Flutter mobile app for YWCA of Bombay, available on PlayStore as "YWCA of Bombay". Enables YWCA to manage events (add/edit/delete), view analytics on user engagement and registrations. Implemented one-click sign-in with Firebase Authentication via OTP and utilized Firebase real-time database for storing event details.

Fullstack Developer

YWCA of India

FlockFlair

Developed "The Botanist" Android app available on PlayStore. Built with Java and XML, utilizes Firebase for user authentication, image storage, and question module. Designed for NEET aspirants.

Android App Developer

Idiot's Room E-Cell, DBIT

Head Of Design

Sequelstring Solutions And Consultancy Pvt Ltd

1. Developed Flutter app for user authorization, utilizing Firebase backend for admin and user authentication. Firebase real-time database used for user authorization.
2. Built analytics dashboard website for Mahindra Auto, developed backend database using Flask Python and Postgres SQL.
3. Implemented ReactJS PWA web app for Pet services and products (customer side) using MERN stack.
4. Developed company website using HTML/CSS and JS.
5. Created Android app for prototyping product listing and graph generation.

Over 2 years of experience in Software Development. Proficient in Python, developing microservices, designing databases, and building APIs. Built an RAG pipeline using OpenAI’s LLM and a vector database. Currently, working with local LLMs and Vector database for enhancing the Generative AI experience.

Taplio

Lav Sharma's Linkedin Analytics

Lav Sharma

Want to drive more opportunities from LinkedIn?