Skip to content
Hardaat Singh Baath

$ whoami

Hardaat Singh Baath

I'm a Machine Learning Engineer at Nurix.AI, building speech and multilingual RAG systems. Before that — deep-learning research, an AI role in Tokyo, and leading autonomy on a student Mars rover.

ML • Systems • Robotics

Projects

Selected work

A few things I've built across ML, robotics, and backend systems — from research to production.

Speech Emotion Recognition

Released

A multi-corpus emotion-recognition model in PyTorch — trained on IEMOCAP, RAVDESS, and CREMA-D — reaching 87% classification accuracy. Built at Nurix.AI.

87% accuracy3 speech corpora
PyTorchAudioDeep Learning

Multimodal Multilingual RAG

Building

A retrieval-augmented generation pipeline serving multiple languages and modalities at 96% accuracy with a p95 latency of 2.93s. Built at Nurix.AI.

96% accuracyp95 2.93s
RAGLLMsRetrieval

Distributed Deployment Platform

Released

Deploys GitHub repos to the web via AWS S3 + ECS, with a Node.js API and reverse proxy, a Kafka log pipeline, and PostgreSQL for real-time build logs.

one-click deploysreal-time build logs
Node.jsAWSKafkaPostgreSQL

Project Kratos — Mars Rover

Research

Autonomous navigation for a student-built Mars rover: visual servoing with YOLO, PID + RTK-GNSS waypointing to 10 cm accuracy, and point-cloud terrain mapping on a Jetson.

10 cm GPS accuracyautonomous nav
ROSComputer VisionRobotics

Milestones

Accomplishments

A few milestones I'm proud of.

  • 2024

    NXP AIM — National Finalist

    Top 15 of 600+ teams. Built an autonomous RC car with lane following, sign detection, and obstacle avoidance.

  • 2024

    LAM Research India — Stage 2

    Top 100 of 1000+ teams. Designed a Peltier-based cooling system with live temperature monitoring and control.

  • 2024

    University Rover Challenge

    Scored 93/100 with Project Kratos, improving on the previous year's result.

  • 2023

    Best Overall Rover — IRC

    International Rover Challenge, Bengaluru — top honour with Project Kratos.

  • 2023

    Published at CVIP 2023

    First-author paper on deep-learning damage segmentation & restoration of Rajasthani wall murals.

Blogs

Blogs

Write-ups from building things — distributed systems, ML, and the low-level details worth remembering. Latest from dev.to.

New writing lands here automatically — for now, read along on dev.to.Follow

Beyond Code

Things that moved me

Essays and half-formed thoughts — on learning, discipline, music, and the things I sit with away from the screen. Latest from Substack.

SubstackOctober 2026

The Truman Show and the Cave We Choose to Stay In

On comfortable illusions, the truth we avoid, and the courage to leave the cave.…

SubstackJuly 2026

I Am My Own Worst Enemy

Reflections from someone still figuring it out.…

Notes

Paper notes

Summaries of papers I'm reading — mostly ML, retrieval, and systems. A living, cross-linked second brain, written to be re-read.

Reading

On the shelf

What I'm reading across the two halves of the brain — no ratings, no reviews.

Technical

  • System Design Interview
  • Designing Data-Intensive Applications
  • Naked Statistics
  • Computer Systems: A Programmer's Perspective

Beyond Tech

  • Humankind: A Hopeful History
  • Thinking, Fast and Slow
  • Slaughterhouse Five

Timeline

The path so far

The moments make the milestones, and the journey makes the maker.

  1. Nov 2021 – Jul 2025

    B.E. Computer Science

    BITS Pilani, Goa

    Bachelor's in CS. Coursework across ML, OS, networks, and architecture.

  2. May 2023 – May 2024

    Autonomous Subsystem Lead

    Project Kratos

    Led autonomous navigation for a student-built Mars rover.

  3. Jun 2023 – Aug 2023

    Research Intern

    CSIR-CEERI, Pilani

    Deep-learning restoration of Rajasthani wall murals.

  4. May 2024 – Aug 2024

    AI Engineer

    DG Takano, Tokyo

    Automated a hardware testing line with cloud logging.

  5. Jan 2025 – Present

    Machine Learning Engineer

    Nurix.AI, Bangalore

    Speech emotion recognition and multilingual RAG systems.

Contact

Let's connect

Building something interesting, or want to talk systems and AI? Drop a line — or find me via the links in the sidebar.