Surti Mohammed Bilal Mohammed Ismailbhai

AI/ML Engineer

Modasa, Gujarat, IndiaArtificial Intelligence and Machine Learning
3roles
38skills
2education
5credentials

About

AI/ML Engineer with hands-on experience in Retrieval-Augmented Generation, FastAPI-based ML deployment, Generative AI applications, and production-ready backend development.

Experience

AI/ML Engineer

CrossTab Technologies LLP

Aug 2025 – Present · India

Built production RAG pipelines using LangChain and ChromaDB for natural language querying of medical JSON datasets; designed hybrid retrieval architecture with semantic search and LLM-based summarization; developed FastAPI backend services with Swagger documentation; integrated Azure OpenAI, Google Gemini, and Perplexity with fallback mechanisms; implemented prompt engineering and response constraints to prevent hallucinations; debugged JSON parsing, API authentication, Docker, and vector database indexing issues; collaborated on code reviews and RAG knowledge sharing.

Data Science Intern

Bharat Intern

Oct 2023 – Nov 2023 · Remote

Developed stock market prediction using LSTM, Titanic survival classification with 81% accuracy, and handwritten digit recognition with 96% accuracy using CNN; performed data preprocessing, exploratory data analysis, feature engineering, model training, and evaluation; created visualizations with Matplotlib and Seaborn.

Python Programming Intern

CodSoft

Oct 2023 – Nov 2023 · Remote

Developed Python GUI applications including a To-Do List manager, Calculator, and Password Generator; applied object-oriented programming, error handling, and input validation.

Education

Maulana Mukhtar Ahmad Nadvi Technical Campus (Affiliated to SPPU)

Bachelor of Engineering, Computer Engineering

Sep 2021 – Jul 2025

CGPA: 6.93/10; 7th Semester SGPA: 8.45; 8th Semester SGPA: 8.75; strong improvement trajectory with consistent academic performance in final semesters.

Scholar Academy of Fine Education

Higher Secondary Certificate (HSC), Mathematics

Jun 2019 – May 2021

Scored 79.8% and ranked among the top 7 students in the institute.

Skills

PythonSQLHTMLCSSScikit-learnTensorFlowKerasPyTorchCNNRNNBi-RNNLSTMNeural NetworksLangChainLangflowRAGPrompt EngineeringVector DatabasesChromaDBQdrant CloudFastAPIREST APIsSwagger UIMicroservicesPandasNumPyMatplotlibSeabornFeature EngineeringMongoDBDockerGitGitHubPostmanCloud ComputingAzureRailwayRender

Projects

Python Programming Q&A Assistant (RAG System)

Built a Retrieval-Augmented Generation Python programming assistant using LangChain, Qdrant Cloud, Groq Llama 3.1, and FastAPI. Indexed 50,000 Stack Overflow Python question-answer pairs, implemented HuggingFace embeddings and similarity search, added hallucination prevention through similarity thresholding, and deployed REST APIs on Railway Cloud.

Healthcare RAG Intelligent Query System

Built RAG pipelines for querying healthcare JSON and document datasets using chunking, embeddings, ChromaDB, LangChain, and Azure OpenAI, with context-aware retrieval and healthcare-specific prompt engineering.

GEO Multi-Endpoint Generative Platform

Developed and tested AI backend endpoints for brand analysis, prompts, reports, and content workflows using FastAPI, Docker, Render, and multi-LLM integrations; optimized API performance and deployment pipelines.

Supplier Information Extraction Chatbot

Built a chatbot workflow for extracting supplier company details, contacts, websites, and locations from unstructured vendor sources using NLP-based entity recognition and structured data extraction.

Shopify Commerce Dataset Analysis

Performed duplicate revenue correction, SKU normalization, mapping validation, and customer-product behavioral analysis on structured commerce datasets; implemented data quality assurance and analytics workflows.

Weather Forecasting using Machine Learning & Deep Learning

Developed comparative forecasting models using CNN, RNN, Bi-RNN, LSTM, and Random Forest on historical weather datasets; implemented sliding-window time-series methods, dropout regularization, and batch normalization; achieved 85.54% overall prediction accuracy across weather categories and models.