Aditya Pandhari

Machine Learning EngineerSingapore

Engineering AI and data systems for real analytical work.

From data engineering to agentic AI. I build analytical systems with clear execution boundaries, dependable data foundations and evaluation built in.

Explore selected workStart a conversation

Working across the analytical system

Three connected areas, supported by validation, documentation, stakeholder communication and technical delivery.

  1. AI Systems and Agentic Engineering

    Designing request handling, analytical workflows, execution boundaries and feedback-led improvements for AI-assisted systems.

    In practiceAnalytical agent for governed data exploration

  2. Data and Analytics Engineering

    Placing data logic in the right layer, building scheduled workflows and shaping analytical products around practical use.

    In practiceAnalytical computation for interactive performance

  3. Applied Machine Learning and Research

    Applying computational methods to text and geospatial questions while keeping methodology, attribution and limitations visible.

    In practiceMovie recommendationFlood-risk mapping

Selected work

Selected engineering problems

Selected work across agentic AI, data engineering and applied machine learning.

  1. ProfessionalAgentic AIJohnson & JohnsonSanitized case note

    Architecting an analytical agent for governed data exploration

    Engineering question How should interpretation, analytical logic and execution controls be separated instead of relying on one unconstrained prompt-to-answer step?

    Problem
    Natural-language analytical requests varied in intent and execution needs, making a single generation path difficult to control consistently.
    Contribution
    Architected the Data Analyst agent, translating requests into SQL, Python and visualization workflows through intent-based routing, shared analytical definitions, query validation and containerised execution, then iterating with stakeholder feedback.
    Stack
    • LangGraph
    • Python
    • SQL
    • Docker

  2. ProfessionalData engineeringSanitized case note

    Re-architecting analytical computation for interactive performance

    Engineering question Which calculations truly require request-time execution, and which can be computed earlier and owned by the data layer?

    Problem
    Recurring analytical logic was being recomputed during user requests, coupling heavy processing to an interactive reporting path.
    Contribution
    Re-architected the workflow by moving repeatable computation into scheduled backend SQL processing and persisting monthly outputs for reporting, comparison and downstream reuse.
    Stack
    • SQL
    • Redshift
    • Airflow
    • React

  3. AcademicFinal Year ProjectApplied ML

    Personalized Movie Recommendation System

    Research focus How can personalization be balanced with discovery, and can a weighted hybrid model capture the strengths of both a classical and a deep-learning approach?

    Problem
    A solo final-year project comparing five recommendation approaches under the tension between personalization and discovery.
    Contribution
    Built and compared an SVD baseline, SVD++, a PyTorch stacked autoencoder, a from-scratch Gaussian RBM and a weighted hybrid model, then validated the hybrid with a 10-person user study.
    Stack
    • PyTorch
    • Surprise
    • pandas
    • Python

    Read the case study Related education: BSc Computer Science, Machine Learning and AI

Explore all projects

Also in the index: Flood Risk Vulnerability Mapping, Trincomalee & Galle

Experience

Progressing from data foundations to AI systems

Data engineering foundations. Increasing responsibility across analytical platforms and agentic AI.

  1. to

    Data Engineer

    Tata Consultancy Services · Johnson & Johnson

    Focus
    Analytical platforms, backend computation and decision-support systems.
    Primary contribution
    Re-architected recurring analytical computation into scheduled backend SQL workflows, persisted reusable outputs for downstream reporting, and engineered data and access logic supporting analytical products.
    Broader scope
    Also worked across forecasting workflows, QA frameworks, reporting modernization, technical documentation and stakeholder-facing delivery.

  2. to presentCurrent

    Machine Learning Engineer

    Tata Consultancy Services · Johnson & Johnson

    Focus
    Agentic analytics, applied AI systems and analytical workflow design.
    Primary contribution
    Architected the Data Analyst agent within a broader multi-agent pilot, designing request interpretation, intent-based routing, shared analytical definitions, query validation and controlled SQL/Python execution.
    Broader scope
    Worked across iterative agent refinement, analytical consistency, stakeholder feedback and evaluation, and internal technical tooling.

Journey

Work is one part of the story.

Education, research, leadership and volunteering have shaped a path that extends beyond professional engineering.

Education

Leadership

SIM Information Technology Club

Progressed from subcommittee member to President of the SIM Information Technology Club, taking on broader responsibility for student leadership and club contribution.

2020–21 Subcommittee
2022–23 President

Recognition

Beyond engineering

Formula 1

Inside a Formula 1 race weekend

Being selected to support a Formula 1 race weekend gave me a very different view of an event I had previously experienced only from the outside. Trackside, the scale comes from the people behind it: marshals, race officials, safety teams and operations working across the circuit.

Aditya Pandhari trackside during a Formula 1 race weekend
Trackside during the race weekend.
A Formula 1 team working together trackside
Teams working across the circuit bring the race weekend together.
Formula 1 race control area
Race control provides the operational view of the circuit.
Formula 1 safety cars on the track
Safety teams are part of the visible rhythm of a race weekend.
Watch from trackside 2 clips
Racing from trackside.
A closing view of the race weekend.

Selected credentials

Learning that follows the work.

A short selection of the certifications and courses I’ve completed while moving from data engineering into machine learning.

Claude certification progression

    Product & delivery

      More learning

        The full archive includes Google’s ML and cloud programmes, earlier coursework and the rest of my verified certificates.

        See the full archive

        Interested in the decisions behind the system.

        I started in data engineering, working closer to the calculations, workflows and reporting foundations behind analytical products. That experience shaped how I approach machine learning and AI today: as systems whose behaviour depends on the data, execution paths, validation and decisions around them.

        I now work as a Machine Learning Engineer while pursuing a part-time MSc in Data Science for Sustainability at NUS. Across professional and academic work, I am most interested in problems that require both technical implementation and judgement: deciding where logic belongs, how a system should be evaluated, and what its outputs genuinely support.

        Outside work and study, my interests have taken me from student leadership to volunteering inside a Formula 1 race weekend. I value opportunities that expose me to different people, disciplines and ways of working.

        LinkedIn GitHub Resume Contact

        Contact

        Let’s discuss thoughtful engineering work.

        Open to conversations about Machine Learning Engineering, AI Engineering, Agentic AI, Data Engineering and related technical work.

        Email AdityaLinkedIn

        A tailored résumé, on request.

        The portfolio provides the broader picture. If you would like a concise résumé for a role, conversation or opportunity, I am happy to share one directly.

        Request résumé by email

        Or email [email protected] with the role or opportunity.