ENGINE: GSAP 3.12 FPS: 60 TWEENS: 0 PUNE: 18.52°N, 73.85°E
 LOC: PUNE, MAHARASHTRA, INDIA (18.5204° N, 73.8567° E) IDENTITY: FULL-STACK SOFTWARE ENGINEER & AI BUILDER
MCA '26 | +91 8830658827 TEJASHREEDUMASIA.ME
00 // IDENTITY — FULL-STACK SYSTEMS & APPLIED AI ARCHITECTURE

Tejashree Vijay Dumasia

Full-Stack Software Engineer & AI Builder

I build software systems, engineer real-time applications, and integrate AI into practical workflows.

Explore Engineered Systems
01 // PROFILE

Architectural Thesis

Software engineering is the discipline of turning complex operational workflows into deterministic, low-latency, production-grade systems.

Trained in Computer Applications—completing my Master of Computer Applications (MCA, 2024–2026) at PES Modern College of Engineering following a 9.16 CGPA in BBA-CA (2021–2024)—I build software from the database schema and API layer upward.

Whether writing C++ for algorithmic problem-solving, architecting REST and WebSocket services in Node.js, TypeScript, and Python, or modeling relational data in PostgreSQL and MySQL, I treat frontend interfaces and backend infrastructure as a single continuous system.

My engineering work focuses on practical system challenges: low-latency WebRTC and Socket.io real-time communication, role-based enterprise applications, and integrating AI/LLMs and computer vision (LangChain, Whisper, OpenCV, EasyOCR, Gemini API) into automated data pipelines.

Rather than treating AI as a novelty, I use structured function calling, signal preprocessing, and strict JSON validation to eliminate repetitive manual operations and make complex databases accessible.

ENTERPRISE CRM SCALE
0 + RECORDS
JWT RBAC PROTECTED
AUTOMATION EFFICIENCY
0 + HRS/WK
MANUAL ENTRY ELIMINATED
UNDERGRADUATE DISTINCTION
0.00 / 10 CGPA
BBA-CA (2021–2024)
PUBLISHED RESEARCH
0 IJSREM DOI
EXPLAINABLE AI IN SECURITY
PILLAR // 01.A

Backend & Relational Architecture

Normalized PostgreSQL and MySQL schemas, JWT role-based access control, and clean REST/FastAPI controllers built for predictable data integrity.

ACID + JWT RBAC POSTGRES / MYSQL
PILLAR // 01.B

Real-Time Communication Systems

Bi-directional WebSocket state synchronization with Socket.io and peer-to-peer WebRTC video/audio conferencing with sub-50ms signaling.

WSS + ICE/STUN WEBRTC / SOCKET.IO
PILLAR // 01.C

Applied AI & Automation Pipelines

End-to-end orchestration combining FFmpeg, OpenCV, Whisper, EasyOCR, LangChain ReAct agents, and Gemini APIs for structured JSON extraction.

LLM DAG + OCR LANGCHAIN / GEMINI
02 // TENURE

Professional Experience

ORGANIZATION // INDUSTRY TENURE

Trinetra Security System

Software Engineering Intern

DURATION: FEBRUARY 2026 – AUGUST 2026 LOCATION: PUNE, MAHARASHTRA
03 // SYSTEMS

Selected Projects

SYSTEM 02 // VISION & LLM EXTRACTION PYTHON • OPENCV • PYTORCH • EASYOCR • GEMINI API

AI Document OCR & Data Extraction Pipeline

An automated document extraction pipeline that processes scanned documents and converts unstructured text into structured JSON.

Engineering Highlights

  • Image Processing using OpenCV: Adaptive thresholding and denoising designed to handle imperfect and blurry document scans.
  • OCR using EasyOCR: PyTorch-backed text detection and bounding-box coordinate extraction.
  • Gemini API & Prompt Engineering: Intelligent data extraction formatted into dynamic, schema-validated JSON output.
Python OpenCV PyTorch EasyOCR Gemini API
GSAP 4-STAGE OCR PIPELINE SIMULATOR CONFIDENCE: 98.8%
RAW SCAN // OPENCV + EASYOCR
INV_NO: #TRN-2026-0891
VENDOR: PUNE INDUSTRIAL LABS
TOTAL_INR: Rs. 1,42,800.00
STATUS: VERIFIED_OK
{
  "pipeline_stage": "READY",
  "opencv_denoise": true,
  "easyocr_tensors": 4,
  "extracted_payload": {
    "invoice_id": "TRN-2026-0891",
    "vendor": "Pune Industrial Labs",
    "amount_inr": 142800.00,
    "verified": true
  }
}
SYSTEM 03 // AGENTIC SQL & INTERNAL TOOLING NODE.JS • LANGCHAIN • SQL • LLM APIs

Natural Language Database Agent

An internal tool that translates natural-language questions into executable SQL queries and natural-language summaries.

Engineering Highlights

  • LangChain ReAct Agent Architecture: Autonomous Thought → Action → Observation loop in Node.js.
  • Dynamic Database Schema Inspection: Custom tools using LLM function calling to inspect table schemas before querying.
  • Safe Query Execution & Summaries: Read-only SQL execution paired with natural-language summaries of database results.
Node.js LangChain SQL LLM APIs Function Calling
GSAP MOTIONPATH REACT SQL SIMULATOR GUARD: READ_ONLY_SAFE
01. THOUGHT Parse Intent 02. TOOL CALL inspect_schema 03. OBSERVATION execute_safe_sql 04. SUMMARY NL Insight
-- PROMPT: "Compare Q1 enterprise subscription revenue across Pune and Mumbai"
SELECT region, COUNT(account_id) AS accounts, SUM(arr_inr) AS total_inr
FROM enterprise_subscriptions
WHERE quarter = '2026-Q1' AND status = 'ACTIVE'
GROUP BY region ORDER BY total_inr DESC;
REGIONACCOUNTSTOTAL_INRINSIGHT
Pune19₹48,20,000Top ARR Cluster
Mumbai14₹36,90,000High Expansion
04 // MATRIX

Technical Skills

GSAP FLIP LAYER FILTER:
RELATIONAL GRAPH: Click any technology node below to highlight companion tools & production deployments. GSAP FLIP READY
LAYER // 04.1

Frontend

LAYER // 04.2

Backend

LAYER // 04.3

Databases

LAYER // 04.4

AI & Data

LAYER // 04.5

Cloud & Tools

LAYER // 04.6

Fundamentals

05 // CREDENTIALS

Academic & Research

05.A // UNIVERSITY DEGREES PUNE, MAHARASHTRA
2024 – 2026 POSTGRADUATE DEGREE

Master of Computer Applications (MCA)

PES Modern College of Engineering, Pune

CGPA: 7.48

Relevant Coursework: Artificial Intelligence, Machine Learning, Deep Learning, NLP & Generative AI.

2021 – 2024 UNDERGRADUATE DEGREE

Bachelor of Business Administration in Computer Applications (BBA-CA)

PES Modern College of ASC, Pune

CGPA: 9.16

Foundation: Relational Database Management, C++ Programming, Data Structures, Web Technologies, and Software Engineering.

05.B // PEER-REVIEWED PUBLICATION VOLUME 10, ISSUE 5 — MAY 2026
INTERNATIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT (IJSREM)

“The Role of Explainable AI in Security Systems: A Survey of Transparent Threat Detection”

Published in IJSREM, Volume 10, Issue 5, May 2026. Authored by Tejashree Vijay Dumasia.

DIGITAL OBJECT IDENTIFIER DOI: 10.55041/IJSREM64095
// BIBTEX FORMAT
@article{Dumasia2026ExplainableAI,
  author  = {Tejashree Vijay Dumasia},
  title   = {The Role of Explainable AI in Security Systems: A Survey of Transparent Threat Detection},
  journal = {International Journal of Scientific Research in Engineering and Management (IJSREM)},
  volume  = {10},
  number  = {5},
  month   = {May},
  year    = {2026},
  doi     = {10.55041/IJSREM64095}
}

// IEEE FORMAT
T. V. Dumasia, "The Role of Explainable AI in Security Systems: A Survey of Transparent Threat Detection," Int. J. Sci. Res. Eng. Manag. (IJSREM), vol. 10, no. 5, May 2026, doi: 10.55041/IJSREM64095.
05.C // TECHNICAL CERTIFICATIONS 04 VERIFIED CREDENTIALS
DATACAMP

Associate AI Engineer for Developers

DATACAMP

Associate Data Scientist

FRONTEND MASTERS

ReactJS

UDEMY

Python for Machine Learning

06 // DISPATCH

Initiate Contact

DIRECT ENGINEERING DISPATCH

Open to full-stack software engineering roles, backend architecture, real-time systems, and applied AI engineering opportunities.

tej.dumasia@gmail.com