●AdityaFileViewWindow
AI Engineer
Engineering Overview

AI engineer building complete systems.

Aditya works across agent architecture, Python and SQL execution, enterprise data platforms, machine learning, validation, and stakeholder delivery. His strength is not a single framework or project. It is the ability to carry technical work across the full system.

01

Systems

Agent platforms, analytics systems, pipelines, and decision-support products.

02

Engineering

Architecture, implementation, controls, testing, and operational delivery.

03

Workbench

How ambiguous problems are framed, decomposed, validated, and improved.

04

Projects

Concrete work across agentic AI, data engineering, NLP, and geospatial analysis.

On mobile, apps open as stacked panels rather than draggable windows.

Systems

Systems built across the stack

Each system combines multiple disciplines rather than isolating one technical feature.

Multi-agent AI platform

Supervisor-led orchestration, specialist agents, governed analysis, Python execution, safety controls, and pilot feedback.

Commercial analytics platform

Redshift transformations, business metrics, Airflow scheduling, regression and trend analysis, and multi-market delivery.

Leadership BI systems

Relational models, filtering, row-level security, and dashboards used by regional heads and client leadership.

Applied ML workflows

NLP distress classification and geospatial flood-risk modelling with domain-specific feature construction.

Engineering

Engineering range

Repeated evidence across systems matters more than a static skill list.

Architecture

System decomposition, specialist-agent boundaries, semantic models, data relationships, and execution paths.

Implementation

Python, SQL, Docker, Redshift, Airflow, YAML, automated testing, and visualisation.

Controls

SQL safety checks, governed metric definitions, row-level security, and source-data validation.

Delivery

Stakeholder clarification, pilot testing, documentation, demonstrations, and informal mentoring.

Workbench

How Aditya works

Recurring engineering behaviours across different systems and domains.

01

Frame the problem

Translate loose business questions, screenshots, metric definitions, or research goals into an implementable system boundary.

02

Model the domain

Define governed metrics, dimensions, relationships, feature sets, and analytical assumptions explicitly.

03

Build the execution path

Connect orchestration, data access, SQL, Python, containers, pipelines, and visual output.

04

Add controls

Introduce validation, query constraints, tests, access rules, and source reconciliation.

05

Validate against reality

Use production-like datasets, reference dashboards, user feedback, and baselines rather than relying on theoretical correctness.

06

Iterate for use

Refine the system through stakeholder sessions, pilot behaviour, and observed operational friction.

Projects

Selected work

Projects are organised by the systems and decisions involved, not by buzzwords.

Enterprise multi-agent analytics

Owned a Data Analyst agent spanning orchestration, semantic resolution, SQL and Python execution, charting, guardrails, testing, and pilot refinement.

Commercial analytics platform

Built Redshift and Airflow pipelines delivering recurring commercial metrics and analytical logic across multiple markets.

Leadership reporting

Delivered Power BI and Tableau systems with relational modelling, filtering, row-level security, and executive consumption.

Dashboard reverse-engineering

Recovered table grain, joins, filters, and metric definitions from raw extracts and reference screenshots.

Distress Signal NLP

Contributed preprocessing, baseline-model comparison, and trend visualisation across roughly 100,000 Reddit comments.

Coastal flood-risk classification

Combined physical hazards and social vulnerability using Google Earth Engine across a 10 km coastal strip.

Research

Applied research

Academic work completed alongside professional engineering.

Distress Signal Detection in Conflict-Period Social Media

Measured anger, fear, and sadness signals across approximately 100,000 Reddit comments during a conflict period against a pre-conflict baseline. Contributed text preprocessing, baseline-model comparison, and visualisation of distress trends.

Coastal Flood Risk Classification, Sri Lanka

Built a two-stage risk classification combining elevation, terrain, coastal proximity, historical inundation, and social vulnerability. Contributed terrain processing, map production, and written analysis.

Profile

Aditya Pandhari

Machine Learning Engineer · Singapore

CurrentMachine Learning Engineer at Tata Consultancy Services, working on a Johnson & Johnson engagement.
EducationMSc Data Science for Sustainability, National University of Singapore, expected 2027.
FoundationBSc (Hons) Computer Science, Machine Learning and Artificial Intelligence, First Class Honours.
CredentialsClaude Certified Developer, Claude Certified Associate, PSPO I, PSM I.
LinksLinkedIn ↗   GitHub ↗