Modern research and software engineering workspace overlooking Singapore
Master's student at NTU Singapore

Padmakar Pimpale

Software Engineer | Applied AI & ML

I build practical software systems across full-stack engineering, data workflows, and AI-assisted applications. My current focus is reliable AI-generated analytics, workflow lineage, and research-backed systems that can hold up in real-world data environments.

View projects Research focus View resume
1.5+ years AI/ML and backend engineering
NTU Singapore MSIS graduate student
Python + MERN Full-stack, ML, databases, APIs
Seeking research AI systems, data lineage, analytics reliability
About

Developer mindset, research direction, product instincts.

Master's student at NTU Singapore with professional experience across AI/ML products, backend systems, full-stack applications, and research-focused software engineering.

I am currently pursuing MSIS at Nanyang Technological University, Singapore, after working across AI/ML products, backend services, and full-stack web applications.

My work includes browser-based financial document analysis, MERN job-platform development, AI math tutoring, a government building-plan scrutiny engine, and data-science/ML prototypes. I enjoy turning unclear product or research ideas into systems that users can test, review, and improve.

Builds with Python, JavaScript, React, Node.js, Django, Flask, databases, and ML libraries.
Comfortable across APIs, dashboards, AI-assisted workflows, data extraction, and deployment.
Currently looking for research opportunities in AI systems, data lineage, and reliable analytics.
Experience

Engineering work across AI products and production backends.

Backend and AI/ML engineering experience across collaboration automation, GovTech scrutiny systems, product R&D, and full-stack platform development.

Jun 2025 - Dec 2025
Backend Developer Sumanel Ventures Pvt. Ltd. (Chat360)

Developed backend services for Microsoft Teams channel creation and management inside Chat360, automating collaboration workflows and reducing manual setup effort.

  • Built Microsoft Teams Channel Integration from scratch using Microsoft Graph API, Azure AD, OAuth 2.0, and permission-handling flows.
  • Implemented end-to-end message parsing and real-time synchronization between Chat360 and Microsoft Teams.
  • Architected Meta Ads Integration, a Facebook Ads Manager style module inside Chat360 using Meta Graph API v24.0.
  • Implemented ad creation, campaign/adset/creative lifecycle management, lead generation flows, and real-time ad analytics dashboards.
Jun 2024 - Mar 2025
AI/ML Developer and Team Lead Techsharthi Consultancy Services, Hyderabad

Worked on in-house product R&D and ABPAS 3.0, a scrutiny-engine project for Madhya Pradesh, India Government, focused on automating building-plan approval workflows.

  • Developed scalable CAD/DXF parsing pipelines with Python, Pandas, Shapely, and ezdxf for FAR, setbacks, geometry, layers, and compliance validation.
  • Built pre-check modules for DXF layers, geometry, file structure, metadata, and reporting before deeper scrutiny processing.
  • Automated multilingual PDF reports with ReportLab, dynamic font scaling, structured tables, and reusable templates for statistics, FAR, heights, and plot metadata.
  • Integrated ezdxf with SVG rendering for geometry, dimensions, annotations, and adaptive TEXT/MTEXT boundary handling.
  • Connected SQL Server/SQLite rule-fetching modules, monitored logs and validation flows with Mobextrem, and optimized runtime stability.
Jun 2023
Project Intern Physics Wallah Pvt. Ltd.

Built a MERN job portal with Node.js APIs, MongoDB schemas, authentication, authorization, client-side routing, and applicant/recruiter workflows.

2020 - 2024
B.E. Computer Engineering International Institute of Information Technology, Pune

Graduated with a CGPA of 8.5, with coursework across data structures, algorithms, object-oriented programming, operating systems, and database management systems.

Projects

Selected projects with working links and practical scope.

Deployed applications and technical prototypes across financial automation, community logistics, job platforms, AI education, compliance systems, and machine learning.

Financial document automation Live on Vercel

Bank Statement ITR Analyzer

Browser-based bank statement PDF to Excel converter with ITR-oriented review summaries. It extracts transaction rows, supports OCR fallback for scanned PDFs, lets users review/edit rows, and exports multi-sheet Excel reports.

Vite JavaScript PDF parsing OCR Excel export
Live app → GitHub →
Shared delivery platform Live on Vercel

BITEZ – Shared Delivery Platform

A neighborhood shared-delivery platform connecting users in Singapore to organize group food orders and reduce delivery costs. Users can discover nearby runs, reserve items, communicate with group members, and coordinate collection.

React TypeScript Supabase Vercel
Live app → GitHub →
Full-stack platform MERN

Job Portal

MERN web app for applicants and recruiters with persistent sessions, JWT-protected REST APIs, recruiter job management, application status flows, fuzzy search, SOP submission, profile pictures, and resume uploads.

MongoDB Express React Node.js JWT
Live app → GitHub →
AI education product AI/ML

AI Math Tutor

An AI-assisted mathematics learning application providing interactive, step-by-step explanations and guided problem-solving. Built with a web frontend and Python/Flask backend to help students understand mathematical concepts.

Next.js TypeScript Python Flask
Live app →
GovTech automation Scrutiny engine

Automated Building Plan Scrutiny

ABPAS 3.0 scrutiny-engine work for Madhya Pradesh, India Government, focused on CAD/DXF parsing, FAR and setback checks, rule-based compliance validation, geometry rendering, and multilingual PDF reporting.

Python ezdxf Shapely ReportLab SQL Server
Official platform →
Data science prototype ML

Metro Interstate Prediction

Predictive maintenance and analytics project for improving metro-system efficiency and reliability, using Python, scikit-learn, TensorFlow, and Keras for machine-learning and deep-learning experiments.

Python scikit-learn TensorFlow Keras
Live app →
Research

Researching reliable AI-generated analytics under changing data.

Current direction: data lineage, provenance, selective recomputation, and validation of AI-generated analytical conclusions when datasets, schemas, or assumptions evolve.

Current research direction

Lineage-Aware Revalidation of AI-Generated Data Science Workflows Under Evolving Data. The goal is to track dependencies between data sources, transformations, SQL/Python workflows, models, visualizations, and textual claims so an AI analysis can be reused, partially recomputed, fully recomputed, or rejected safely after data changes.

Proposed benchmark: ChangeImpactBench for evaluating workflow validity after row, value, schema, label, and business-rule changes.
Core methods: dependency graphs, provenance tracking, selective recomputation, uncertainty-aware reuse/abstain decisions.
Target domains: finance, healthcare, e-commerce, logistics, and enterprise decision-making where stale analytics create risk.
Research in progress | NTU | 2026

When Should Humans Intervene? Comparing Proactive and Reactive Oversight in Autonomous AI Decision-Making

This research explores how the timing of human intervention influences trust, perceived control, and responsibility in autonomous AI decision-making. It compares proactive oversight, where humans intervene before an AI system acts, with reactive oversight, where humans review decisions after the system has acted.

Ongoing research; no completed experimental results claimed.

Published research | AI floor-plan generation

Automated AI Floor Plan Generation - Paper 1

Prior research work on AI-assisted architectural floor-plan generation and automation in design workflows.

IJNRD2405604 PDF →
Published research | AI floor-plan generation

Automated AI Floor Plan Generation - Paper 2

Related publication connecting generative AI, floor-plan automation, and intelligent support for planning workflows.

IJNRD2403536 PDF →
Research interests

Actively looking for research collaboration

Interested in AI systems, data provenance, reliable LLM agents for analytics, automated data-science workflows, evaluation benchmarks, and applied machine-learning systems.

Skills

Technical stack grouped by how I use it.

Core technologies used for backend services, web applications, AI/ML workflows, data systems, deployment, and research prototypes.

Languages

Python, JavaScript, Java, C++, SQL

Frontend

React, Next.js, Vite, responsive UI, dashboards, client-side routing

Backend

Node.js, Express, Django, Flask, REST APIs, authentication, JWT

Databases

PostgreSQL, MySQL, MongoDB, schema design, query workflows

AI/ML

Pandas, NumPy, Matplotlib, Seaborn, scikit-learn, PyTorch, Keras, TensorFlow

LLM Systems

OpenAI API, Google Gemini, RAG concepts, AI-assisted workflows, evaluation

Tools + Cloud

Docker, Vercel, Netlify, GitHub, Supabase-ready deployments

Foundations

DSA, OOP, operating systems, DBMS, technical writing, research reading

Contact

Open to research, internships, and software engineering opportunities.

I am especially interested in AI systems, reliable analytics, backend/data platforms, and practical products that combine software engineering with research thinking.

2002padmakar@gmail.com → View resume → Download resume → GitHub → LinkedIn → Selected work →