cv

Basics

Name Jayanth S
Label AI/ML Research Specialist
Email jayanths2498@gmail.com
Url https://js2498.github.io
Summary AI/ML Research Engineer with a background in working with deep learning models, building LLM-based retrieval systems, and applying machine learning to wireless communication problems. Hands-on experience with end-to-end ML systems, retrieval pipelines, and experimentation workflows. Current interests include Machine Learning, Stochastic Optimization, Reinforcement Learning, and applied research.

Work

  • Mar 2025 – Present
    6G AI/ML Research Specialist
    Nokia Standards
    AI/ML research for 6G physical layer and LLM-based internal tooling.
    • Built LLM-based retrieval workflows for summarization, question answering, information extraction, and cross-document comparison.
    • Developed RAG pipelines using embeddings and vector search over structured and unstructured technical data.
    • Contributed to React-based frontend and Streamlit prototypes for internal AI tools.
    • Designed multi-step workflows using LangChain and LangGraph with prompt engineering for task routing and automation.
    • Worked on deep learning models for signal processing and multi-antenna communication systems.
    • Migrated and validated a large-scale simulation and ML codebase (~10k LOC, 80 modules) from TensorFlow to PyTorch.
    • Implemented reproducible ML experimentation pipelines using MLflow, Flyte, and Argo Workflows.
  • Feb 2023 – Mar 2025
    Research Engineer
    TCS Research
    • Won the Depth Estimation using mmWave AI/ML Challenge - 2023 organized by NIST (USA) and ITU.
    • Implemented OFDM and OTFS communication systems on USRP B210 SDRs for real-time wireless prototyping.
    • Demonstrated image/data transmission using OFDM over SDR with a RIS-assisted wireless link.
    • Developed cascaded channel estimation algorithms for RIS-aided mmWave systems — IEEE ICC Workshop 2024.
  • Oct 2022 – Dec 2022
    Research Assistant, Dept. of Computer and Information Science
    Linköping University
    • Worked on stochastic control methods for semantic communication efficiency, Age of Information, and cost-aware scheduling.
    • Research led to WiOpt 2023 publication with subsequent conference publication as co-author.
  • May 2022 – Aug 2022
    Research Intern
    TCS Research
    Reinforcement learning for wireless communications.
    • Implemented DDPG, TD3, and SAC RL algorithms for hybrid beamforming in single-user MIMO communication systems.

Education

  • 2020 – 2022

    Dharwad, Karnataka, India

    MS (Research)
    Indian Institute of Technology, Dharwad
    Electrical Engineering — Optimization of Age of Information in Wireless Communication Networks
  • 2015 – 2019

    Bengaluru, Karnataka, India

    B.E.
    PES Institute of Technology, Bangalore South Campus
    Electronics and Communication Engineering

Awards

Certificates

Publications

Skills

Languages
Python
C++
MATLAB
ML Frameworks
PyTorch
TensorFlow
Scikit-learn
Ray RLlib
GenAI Systems / Tooling
LangChain
LangGraph
Streamlit
Elasticsearch
Retrieval-Augmented Generation (RAG)
Prompt Engineering
MLOps / Dev Tools
Git
MLflow
Flyte
Argo Workflows
Relevant Coursework
Reinforcement Learning
Statistical Pattern Recognition
Convex Optimization
Probability and Stochastic Processes

Languages

English
Professional working proficiency
Kannada
Native speaker
Telugu
Conversational

Interests

Research
6G Wireless
LLM Systems
Age of Information
Reinforcement Learning
Stochastic Control

Projects

  • 2026.01 - Present
    nanoQwen — LLM Implementation from Scratch
    • Implemented a Qwen-style decoder-only Transformer from scratch in PyTorch with RoPE, RMSNorm, Grouped-Query Attention (GQA), causal self-attention, and autoregressive text generation.
    • Built end-to-end pretraining and supervised fine-tuning (SFT) pipelines with custom dataset preprocessing, mixed-precision training, checkpointing, configurable experiment workflows, and tokenizer integration.
    • Developed modular training and inference infrastructure including KV-cache based decoding, HuggingFace parity validation, model evaluation utilities, and configurable scaling for transformer experimentation.
  • 2022.01 - Present
    Policy Gradient Algorithms for Atari Games
    • Applied A2C, A3C, TRPO, and PPO based on the stablebaselines3 and ray rllib implementations for Pong, Breakout and Space-Invaders atari games.
  • 2022.01 - Present
    Model-Free End-to-End Communication Systems
    • Implemented the model-based and model-free auto-encoder based end-to-end communication system for AWGN and Rayleigh Block Fading channels as in: Fayçal Ait Aoudia and Jakob Hoydis, "Model-Free Training of End-to-End Communication Systems."
  • 2022.01 - Present
    PrecoderNet — Hybrid Beamforming
    • Implemented deep learning precoder for hybrid beamforming in massive MIMO systems using DDPG, TD3, and SAC reinforcement learning algorithms.