CV

Full CV as PDF via the icon on the right.

Contact Information

Name Vatsal Jha
Professional Title Machine Learning Engineer and Researcher
Email vatsaljha17@gmail.com
Website https://vatsaljha.dev

Professional Summary

I build machine learning systems that make decisions, and the infrastructure that runs them.

Experience

  • 2024 - present

    Vellore / Remote

    Researcher
    Vellore Institute of Technology (with Dr. B. Ashok)
    • Federated learning, bandits for EV charging placement, and selection-aware carbon claims. Co-inventor on Indian Patent Application 202641089937.
  • 2026 - 2026

    Remote

    Machine Learning Intern
    testAIng
    • Multilingual audio-deepfake detection with ECAPA-TDNN, WavLM and mixture-of-experts models: 0.94–0.96 AUC on internal multilingual benchmarks, 1.97% EER after Hindi fine-tuning.
    • Moved inference to Docker/Kubernetes with autoscaling, probes, rolling deploys and Prometheus/Grafana.
  • 2024 - 2025

    Remote

    Research Collaborator, handwritten text recognition
    Remote collaboration with Dr. Sukalpa Chanda and Aniket Anand Gurav
    • End-to-end handwriting detection and recognition on IAM: ResNet-50 backbone, FOTS-style detector and an attention-augmented PHOSC recognizer.
  • 2025 - 2025

    Bangalore, India

    Software Engineering Intern, Server Platform
    Samsung Research India
    • Designed and built a multi-agent root-cause analysis system with Google ADK and MCP over Prometheus, Kubernetes, Jaeger, OpenSearch and Milvus; evaluated on an OpenTelemetry Demo + Chaos Mesh fault-injection testbed.
    • Received a full-time pre-placement offer.
  • 2024 - 2025

    Hybrid

    Team Lead / Co-primary Researcher
    Samsung R&D Institute Bangalore (PRISM)
    • Co-led four student researchers on HOMA, a multilingual LLM multi-agent system for smart homes (11 Indian languages). Samsung PRISM Excellent Research Award (2025).
  • 2024 - 2024

    Delhi, India

    Data Science Intern
    Council on Energy, Environment and Water (CEEW)
    • NLP over ~93K task descriptions from 600+ occupations for job-transition analysis; ML bias correction of Delhi PM2.5 forecasts (IICAQM 2024 poster).
  • 2022 - 2024

    Hybrid

    ML Research Intern
    VIT × King Saud University (with Dr. B. Ashok)
    • Software and methodology for electric two-wheeler state-of-charge prediction from on-road CAN/BMS telemetry; published in the Journal of Energy Storage.

Education

  • 2022 - 2026

    Vellore, India

    B.Tech.
    Vellore Institute of Technology
    Computer Science and Engineering (Internet of Things)
    • NPTEL Reinforcement Learning (IIT Madras): top 1% nationally; invited to interview for Instructor, Reinforcement Learning, IIT Madras BS programme.
    • NPTEL Responsible & Safe AI Systems (IIIT Hyderabad × IIT Madras): top 5% nationally.

Awards

  • 2025
    Best Paper Award, IEEE ICDSINC 2025
    IEEE ICDSINC

    For CA-AMFA.

  • 2024
    National Winner, IISF 2023 Space Hackathon
    ISRO / NRSC Bhuvan
  • 2024
    National Winner, Samsung PRISM Hackathon
    Samsung R&D Institute Bangalore
  • 2026
    1st Prize, testAIng intern research presentation
    testAIng
  • 2025
    Samsung PRISM Excellent Research Award
    Samsung R&D Institute Bangalore

    For HOMA.

  • 2024
    Top-3 Finalist, Karnataka State Police Datathon
    Karnataka State Police

    Out of 380 teams.

  • 2024
    Raman Research Award
    Vellore Institute of Technology

    For the EV state-of-charge research.

Leadership

Team Prometheus, VIT RoboCup Small-Size League (2024–2026): Co-founder and co-lead; grew the team to 30+ students and secured in-kind sponsorship from SolidWorks, ANSYS and Altium.
Youth Red Cross, VIT (2024–2025): Vice Chairperson, then Chairperson; 1,000+ blood donations, including a camp that set VIT's single-day record (750+ units).

Skills

ML: PyTorch, Hugging Face, scikit-learn, Flower, RL and bandits, federated learning, LLM agents, RAG, evaluation
Systems: Docker, Kubernetes, KEDA, FastAPI, Ray Serve, vLLM, MCP, PostgreSQL, Redis, RabbitMQ, Prometheus, Grafana
Languages: Python, Go, C++, Java, SQL, TypeScript