Ansuk Sinha Roy

M.Tech researcher in Communication and Signal Processing at IIT (ISM) Dhanbad, working on the wireless physical layer – online learning for channel estimation and system-level RRM simulation. Previously interned with Nokia's Standards team, applying deep RL to beam management inside their internal simulator, SAMPO.

ANRF Research Fellow at IIT (ISM) Dhanbad (2026-08 – Present)

• Developing an online-learning approach to physical-layer channel estimation under high-Doppler conditions, where wireless channels vary too quickly for standard estimators to track well.

Research Intern at Nokia Technology Standards (2025-08 – 2026-02)

• Tested whether deliberately sub-optimal beam selection could raise network throughput under RRM constraints, evaluating CSI-RS beam selection inside SAMPO – Nokia's internal system-level RAN simulator – against RSRP, beam usage, and inter-beam cross-correlation.
• Replaced the vanilla DQN baseline with a Dueling Double DQN, added Prioritized Experience Replay, and enforced feasibility with constraint-aware action masking – improving throughput 15–35% over the heuristic baseline.
• The approach was validated well enough to be approved for contribution to SAMPO's main repository.

Master of Technology in Communication and Signal Processing – Indian Institute of Technology IIT (ISM), Dhanbad (2024 – 2026)

Bachelor of Technology in Computer Science and Technology – University of Engineering and Management, Kolkata (2020 – 2024)