Anas Sayed | AI Engineer | FinTech & Decision Intelligence

About Me

I am a BSc Computer Science student focused on Artificial Intelligence, Machine Learning, and FinTech, with a particular interest in credit risk, financial data, and decision intelligence.

I build practical AI systems that combine machine learning with real-world software engineering — from explainable credit risk models and financial research agents to production-oriented AI applications and intelligent data-driven tools.

My current work spans Credit Risk, Financial Data Engineering, Generative AI, Deep Learning, and AI-powered decision systems. I am particularly interested in applying AI to financial services, risk, analytics, and other high-impact domains.

Education: B.Sc. Computer Science — D.G. Ruparel College, Mumbai (2024 – 2027)

Certifications:

AI Security Fundamentals – LLM Threats & OWASP 2026 — Packt (In Progress)
Generative AI with Large Language Models — DeepLearning.AI & AWS (Coursera)
Machine Learning Crash Course — Google (2024)
Excel for Business — Great Learning (2024)

Experience

Machine Learning Engineer Trainee (OJT)

Fusion Klassroom Edutech Limited • Mumbai, India • December 2025 – May 2026

Annotated and structured information from video-based sources into Excel and Google Sheets datasets for machine learning workflows.
Performed contextual analysis and inference to identify and populate structured data fields from unstructured video content.
Conducted quality verification and validation of annotated datasets while maintaining accuracy and meeting project deadlines during the On-the-Job Training program.

Featured Projects

CreditSense — AI-Based Credit Risk Scoring System

End-to-end FinTech credit risk scoring system analyzing structured financial transaction records and assessing borrower default probabilities. Evaluated against 307,000+ loan records using a 20-feature financial behavior pipeline, SQL/PostgreSQL data extraction, XGBoost scoring with SHAP explainability, Account Aggregator data processing, and Dockerized FastAPI APIs.
- View Code
StockLens — Deep-Research AI Agents & Fundamental Stock Screener

Autonomous multi-agent equity research and fundamental screening platform for Indian markets (NSE & BSE). Combines deep-research workflows, grounded financial-data retrieval, 11-factor fundamental analysis, Graham valuation, multi-scenario DCF, and resilient AI model orchestration.
- View Code
Multi-Crop Plant Disease Detection & Farmer Advisory System

Academic mini-project implementing an end-to-end crop diagnosis and farmer advisory application using a Dual-Head Vision Transformer (ViT) architecture for simultaneous classification across 55+ crop species and 170+ diseases. Built with MobileNetV2 benchmark comparisons, OWASP-reviewed JWT authentication, SQLite storage with structured SQL queries, and FastAPI-based inference.
- View Code
Career Funnel Optimizer — AI Resume & ATS Optimization Platform

Full-stack AI-powered resume optimization platform that analyzes job descriptions, evaluates resume–job fit, and generates ATS-compatible resume modifications. Uses structured LLM outputs, hybrid scoring, deterministic anti-hallucination validation, and targeted DOCX editing while preserving document formatting.
GoalsApp — 12-Week Goal Execution & Habit Tracking System

Production-hardened offline-first mobile application for structured goal execution, habit tracking, daily pacing, and spaced-repetition learning. Includes hierarchical 12-week goals, dynamic daily pace calculation, consistency heatmaps, revision scheduling, atomic storage operations, and automated Android backups.
- View Code

Engineered From Scratch

Every model, data pipeline, and retrieval layer below was designed, coded, and validated end-to-end — focusing on deep conceptual mastery of core architectures and production workflows.

DevDocs AI — RAG-Based Documentation Assistant

Retrieval-Augmented Generation application for question answering over technical documentation. Built the retrieval layer from first principles using Hugging Face embeddings and custom NumPy cosine-similarity search, then orchestrated retrieved context into LLM prompts for grounded responses.
- View Code
Enterprise AI Credit Intelligence & Macro-Risk Engine

Unified decision intelligence engine utilizing a Parsimonious 5-Feature Architecture for credit scoring. Implements XGBoost with SHAP explainability, automated macro ETL (NIFTY 50, GDP, Inflation via yfinance and fredapi), ADF time-series stationarity testing, and a Dockerized FastAPI microservice with fast response times.
- View Code
SubSlash AI — Autonomous FinTech Marketing Agency

Autonomous multi-agent system built with CrewAI and Google Gemini that autonomously plans, researches, and executes a full Go-To-Market campaign. Orchestrates 4 specialized AI agents (Strategist, Researcher, Viral Creator, SEO Writer) for automated competitor gap analysis, content calendars, and scriptwriting without human intervention.
- View Code
Telco Customer Churn Prediction

End-to-end supervised classification pipeline predicting customer churn on 7,000+ telecom accounts. Handled data preprocessing, One-Hot & Label encoding, and feature scaling, evaluating Logistic Regression, Random Forest, and XGBoost with decision threshold tuning optimized for maximum churn recall.
- View Code
Used Car Price Regressor

Supervised regression pipeline predicting secondary market vehicle valuations across 6,000+ vehicle records. Executed exploratory data analysis, derived engineered temporal features (car_age), and evaluated against an initial Ordinary Least Squares baseline ($R^2 = 0.54$), achieving an improved $R^2$ of 0.69 via a non-linear Random Forest Regressor.
- View Code