👋 Hello, I'm

Shubham Sapkal

I'm a |

Software Engineer specializing in Python, Django & REST APIs, with applied AI experience across RAG/LLM systems and Computer Vision pipelines — building distributed, containerized backend systems that scale.

View My Work Hire Me
01. About

Who Am I?

Shubham Sapkal
Backend & Applied AI Engineer

Software Engineer — Backend & Applied AI

Software Engineer with 1.3+ years of experience building REST APIs, event-driven backend services, and distributed systems in Python and Django. Skilled in containerized deployment (Docker, Nginx), asynchronous messaging (RabbitMQ), and applied AI — including computer-vision pipelines (YOLOv8, OpenCV) and offline Retrieval-Augmented Generation systems (Sentence Transformers, Ollama, vector search).

I delivered a fully offline document-intelligence system demonstrated live to enterprise stakeholders at the HP AI Experience Center. I love turning complex, ambiguous problems into elegant, production-ready systems.

1.3+
Years Exp.
4+
Projects
15+
Technologies
Get In Touch GitHub
02. Experience

Work History

Software Engineer

Augmented Transformations Pvt Ltd
May 2025 – Present
  • Architected and delivered RESTful APIs and reusable Python/Django modules across multiple application modules, cutting feature-delivery time by ~30%.
  • Diagnosed and resolved a critical API performance bottleneck by redesigning query structure and database interaction patterns, reducing response time from 12s to 2s (83% improvement).
  • Standardized Docker-based deployment workflows and Nginx reverse-proxy configuration across development and production environments, cutting environment-setup time by ~70%.
  • Optimized SQL queries and indexing strategy for large-scale operational datasets, improving retrieval performance by ~40% during high-volume processing.
  • Automated recurring data-processing workflows and backend utilities, cutting manual operational effort by ~50%.
  • Trained and fine-tuned OpenCV/YOLOv8 computer-vision pipelines for dataset preparation, model training, and evaluation, improving model accuracy by 12–15%.
  • Built and deployed a fully offline, local-LLM document-intelligence pipeline (Sentence Transformers + Ollama + vector search) with no external API dependency — demonstrated live at the HP AI Experience Center.
PythonDjangoDRFDockerNginxRabbitMQMySQLYOLOv8OpenCVOllamaRAG
03. Skills

Technical Arsenal

Languages
PythonSQLJavaC
Backend & APIs
DjangoDjango REST FrameworkRESTful API DesignAuth & Authorization
Databases
MySQLFirebase Realtime DBQuery OptimizationIndexing
Distributed & Real-Time Systems
RabbitMQWebSocketsEvent-Driven Arch.Async Processing
AI / LLM Engineering
RAGSentence TransformersOllamaVector SearchSemantic Retrieval
Computer Vision
OpenCVYOLOv8Model TrainingFine-TuningEvaluation
DevOps & Infra
DockerNginxGitLinuxContainerization
Data Processing
PandasNumPyETL Pipelines
Frontend
AngularHTMLCSS
04. Projects

Featured Work

01

SenitelAI

Edge-Based Traffic Violation Detection System

  • Developed an edge-based traffic-monitoring and violation-detection platform processing live video streams from surveillance cameras.
  • Deployed vehicle detection, tracking, and traffic-event analysis pipelines using YOLOv8 and OpenCV for real-time violation identification.
  • Established distributed, event-driven workflows using RabbitMQ and shipped Django REST APIs with low-latency pipelines handling concurrent events and high-volume streaming data.
PythonDjangoAngularYOLOv8OpenCVRabbitMQGoDocker
Code
02

CuratorAI

Offline RAG Document Intelligence System

  • Constructed an end-to-end RAG pipeline — ingestion, chunking, embedding generation, vector indexing, and local LLM generation via Ollama — for offline semantic document search and Q&A.
  • Established scalable backend workflows to process large volumes of unstructured documents while maintaining efficient retrieval performance, with zero external API dependency.
  • Enabled semantic search, contextual retrieval, and automated knowledge extraction using vector search and RAG architecture principles.
PythonSentence TransformersOllamaVector Search
Code
03

PlaceSync

Real-Time Campus Recruitment Platform

  • Developed a real-time recruitment platform with admin and student interfaces.
  • Enabled live updates using Firebase, reducing manual coordination by 80%.
  • Designed dynamic UI components using RecyclerView (Android).
AndroidJavaFirebase Realtime DBRecyclerView
Code Student Code Admin
04

Twwitt Karo

A Mini Twitter Clone (Django + SQLite)

  • Built a full stack social media app with authentication and full CRUD features.
  • Designed responsive frontend and efficient backend routing with Django.
  • Improved usability through optimized architecture.
DjangoPythonSQLiteHTML/CSS
Code
05. Certifications

Certifications

Python Django

Simplilearn

Issued Dec 2025 · Credential ID: 9582173

DjangoDjango REST Framework
Python For Beginners

Simplilearn

Issued Dec 2025 · Credential ID: 9581945

PythonNumPy
Developing SQL Databases

Simplilearn

Issued Nov 2025 · Credential ID: 9407887

PostgreSQLDatabases
06. Education

Academic Background

Bachelor of Computer Application

BCA · 8.54 CGPA

MES Abasaheb Garware College Of Science

Savitribai Phule Pune University

April 2025
07. Contact

Let's Work Together

Get In Touch

Whether you're looking for a backend or applied-AI engineer, want to discuss a project, or just want to talk tech — I'm always open to new opportunities!

Phone

+91 9373290652

Location

Pune, Maharashtra, India