Languages
Python, SQL, JavaScript, TypeScript
Software Engineer [Noida, India]
I build reliable, thoughtful software that solves real problems.
I’m Suraj Singh, a Software Engineer engineer focused on Artificial Intelligence. I enjoy turning complex product needs into clear, durable experiences.
* Start * Learn * Build * Fail * Repeat
About
Building with care, from first idea to lasting impact.
I'm a Software Engineer living in Noida, India. I work as a Software Engineer at Capgemini. where I help build innovative solutions. currently Working with PyTorch, RAG, LLM. I'm interested in the gap between a model that works in production data pipelines, evaluation, latency, and failure modes included.
I focus on making models measurable before I call them done clear eval metrics, reproducible experiments, and monitoring once something ships.
Capabilities
Tools for thoughtful delivery.
Frontend
HTML5, CSS3, JavaScript, React, Data Visualization
Backend
FastAPI, Flask, Node.js, API Integration, Backend Automation
Cloud & DevOps
AWS, Docker, Git, Github, Lambda, Cloud Deployment
Databases
MySQL, PostgreSQL, MongoDB, Data Processing
Tools
Pandas, NumPy, Sckit-learn, TensorFlow, PyTorch, OpenAI APIs, Gemini API, Power BI, VS Code, Cursor
Experience
A record of shipping, learning, and improving.
Capgemini
Software Engineer
- Built AI-powered workflows with prompt engineering and enterprise data sources, automating recurring business processes and cutting manual document/knowledge lookup time for internal teams.
- Partnered with cross-functional engineering teams to design AI solutions combining LLMs, backend services, and Azure cloud architecture, taking features from design through to production deployment.
- Redesigned citation and source-attribution logic for AI-generated responses, reducing unsupported/unverifiable outputs and increasing stakeholder trust in the system's answers.
MNG Realtech
Data Analyst
- Built Power BI/Tableau dashboard to track occupancy, revenue, and sales performance,improving reporting turnaround time by 28%.
- Built Predictive models for property demand and pricing trends.
- Cleaned and validate large customer datasets for data quality and setup cloud-connected data pipelines for scalable analysis.
Selected work
Featured projects.
01
Project 01
Enterprise Operations AI Agent
An agentic AI assistant that investigates enterprise operational issues such as blocked orders, payment failures, credit-limit violations, and inventory problems by reasoning over business data and company policies.
Problem: Enterprise operations teams often spend significant time manually checking orders, customer information, payment history, and business policies to identify why an order is blocked. This project automates the investigation process while keeping the reasoning explainable and evidence-based.
Technical decisions & challenges: Built a Python/FastAPI backend with SQLite and SQLAlchemy, integrating Gemini for AI reasoning and function/tool calling. Designed modular tools for retrieving orders, customer information, payment history, and company policies. The main challenge was creating a controlled workflow where the AI can decide which tools to use without inventing enterprise data.
Outcome: Built a working foundation for an agentic enterprise investigation system that can retrieve real database information, analyze operational issues, identify potential root causes, and provide evidence-based recommendations. The architecture is designed to evolve towards LangGraph workflows and human-approved business actions.
Tech: Python, Gemini API, FastAPI, SQLAlchemy, SQLite Tool Calling, React
02
Project 02
IPL Score Predictor
Summary: A machine learning web app that predicts the final score of an IPL first innings in real time, based on the current match state — Batting team, Bowling team, Runs scored, Wickets lost, Overs left, and momentum in the last 5 overs.
Problem: Fans, analysts, and commentators constantly estimate where an innings will finish, but that guesswork is usually gut-feel rather than data-driven. This project turns live match variables into a statistically grounded score projection updated dynamically as the innings unfolds.
Technical decisions & challenges: Built a Python backend using Flask/FastAPI, with an XGBoost regression model trained on historical IPL ball-by-ball data. Engineered features capturing mid-innings momentum without overfitting to any single team or venue, maintaining stable predictions early in match play.
Outcome: Built and deployed a live predictor service that delivers projected final scores with robust accuracy across diverse match situations.
Tech: Python, Scikit-Learn, XGBoost, Flask, FastAPI, Pandas, NumPy, React
03
Project 03
FITGENIE (Personal AI GYM Trainer)
Summary: An AI-powered personalized fitness web app that generates tailored workout and nutrition guidance based on user's profile including age, height, weight, fitness level, goal, and dietary preferences while also helping users track their progress and stay motivated.
Problem: Most fitness plans follow a one-size-fits-all approach and don't adapt to an individual's goals, fitness level, or dietary prefernces. FitGenie turns personal fitness information into actionable workout plans and fitness guidance through an interactive AI-powered experience.
Technical decisions & challenges: Built a responsive full-stack web application using React and Python, integrating AI/API capabilities to generate personalized fitness recommendations. Engineered structured prompts and user inputs to produce relevant workout, nutrition, and fitness guidance while keeping the experience simple and interactive. Added weight progress loging with measurements, dates, and notes to help users monitor their progress over time.
Outcome: Built and deployed a live AI-powered fitness assistant that delivers personalized workout and nutrition recommendations, fitness tips, and progress tracking through a single web application.
Tech: Python, React, AI/API Integration, Prompt Engineering, Vercel, Full-Stack Development
Education
Education
Bachelors in Technology (CSE AI Honours)
Noida Institute Of Engineering And Technology, Gr. Noida• 2020-2024