Profile

Abhinav Sharma

Not just models. Not just Code. I build complete systems

About Me

I'm most interested in the part of the process where things are still unclear, where the path ot a solution isn't obvious. I work at that intersection of data and systems, building everything from pipelines and models to applications that people can actually use. Because in the end, what matters is not whether something works in theory, but whether it works in the real world.

With a strong foundation in both technical and analytical skills, I bridge the gap between data science and software engineering to deliver comprehensive solutions that drive business value.

Skills & Technologies

Technical expertise spanning data science, machine learning, and full-stack development

Data Science & Analytics

Python Expert
R Familiar
SQL Expert
Pandas/NumPy Expert
Scikit-learn Proficient
TensorFlow/PyTorch Proficient
Jupyter Expert
Tableau/Power BI Proficient

Software Engineering

JavaScript/TypeScript Expert
React/Next.js Expert
Node.js Proficient
Git/GitHub Expert
Docker Proficient
AWS/Cloud Proficient
PostgreSQL/MongoDB Proficient
REST APIs Expert

Machine Learning & AI

Statistical Analysis Expert
Data Visualization Expert
Predictive Modeling Proficient
Deep Learning Proficient
Expert - Advanced proficiency
Proficient - Strong working knowledge
Familiar - Basic understanding

Featured Projects

A showcase of my data science and software engineering work

ReviewShieild - Fake Review Detection API

Fake reviews are easy to generate, but harder to detect reliably in a way that can be used in real systems. The goal here was not just to train a model, but to build something that could actually be integrated and used. I designed an end-to-end system that takes raw text input and returns a fraud likelihood score through a clean API. The focus was on making the system usable and deployable, not just accurate. This meant thinking beyond modelling — designing the API, handling requests, and ensuring the system runs reliably in a production-like setup.

Python Scikit-learn Flask Docker

ReviewShield – Fake Review Detection API

Detecting fraudulent reviews before they influence real users.

Fake reviews are easy to generate, but much harder to detect in a way that can be used in real systems. This project started as a machine learning problem, but quickly became a broader systems challenge.

I built an end-to-end solution that takes raw review text, processes it through an NLP pipeline, and returns a fraud likelihood score through an API. The focus was not just model performance, but building something practical, usable, and ready to be deployed.

What I built

  • A complete NLP pipeline from preprocessing to prediction
  • An API layer that exposes the model for real-world use
  • A deployable system designed for reliability and continuous availability
Python Scikit-learn FastAPI Uvicorn Nginx Cloudflare

Stock Market Prediction System

Developed a deep learning model for stock price prediction using LSTM networks and technical indicators. Includes backtesting framework and risk analysis.

Python TensorFlow Pandas Streamlit

Work Experience

Research Software Engineer

University of Leeds, UK

2022 - Present

Test Engineer

Infosys, Hyderabad, India

2018 - 2019

Education & Certifications

MSc in Computer Science

University of Leeds, United Kingdom

2021 - 2022

Specialized in machine learning, statistical modeling, and big data analytics. Thesis on data engineering and data science on predicting social unrest events.

Machine Learning Statistical Analysis Data Mining Deep Learning

BTech in Computer Science & Engineering

DIT University, Dehradun, India

2014 - 2018

Strong foundation in algorithms, data structures, and software engineering principles.

Algorithms Data Structures Software Engineering AI