We publish
what we build.

Both co-founders have peer-reviewed publications. Research isn't a marketing prop — it's why clients hire us for the work nobody else will take.

What we're researching now.

Three active research tracks, two submissions under review, one open collaboration.

Area 01

Realtime multiplayer systems

Networking, state reconciliation and anti-cheat for card and strategy games at sub-100ms. In direct service of PokerEdge.

Lead: A. NittalaActive
Area 02

Retrieval for agentic systems

Grounded retrieval and evaluation for long-horizon LLM agents. Published at a NeurIPS workshop; a full-paper submission is under review.

Lead: A. NittalaSubmitted
Area 03

Low-cost agri-sensing

Sensor networks and autonomous field systems for smallholder farms in India. Field trials in Maharashtra and Karnataka, with ICRISAT collaboration.

Lead: P. BawseField trials

Five peer-reviewed papers.

Across ML, realtime systems and HCI. Filter by area below.

2025

Latency-aware move prediction in live multiplayer card systems

A. Nittala, P. Bawse, K. Menon
A prediction layer that speculatively renders card-game state 40ms ahead of server confirmation, reducing perceived latency by 62% in cross-region sessions.
IEEE Games 2025
PDF soon
2024

Retrieval-augmented coaching signals in turn-based strategy games

A. Nittala, P. Bawse
Combines GTO solver output with game-state retrieval to generate natural-language coaching signals, evaluated across 12k human-played hands.
NeurIPS 2024 WS
PDF soon
2024

A low-cost capacitive soil-moisture network for smallholder farms

P. Bawse, A. Nittala
A $9 sensor node and LoRaWAN gateway design deployed across 23 farms in Maharashtra, with a year-long reliability study.
ACM COMPASS
PDF soon
2023

Autonomous trajectory planning for sub-250g aerial systems in cluttered fields

P. Bawse, A. Nittala, R. Desai
A lightweight SLAM + RRT* variant for sub-250g drones in densely planted agricultural fields, with outdoor evaluation across 3 crop types.
IROS 2023
PDF soon
2023

Constrained-domain fine-tuning for low-resource Indic language pairs

A. Nittala, N. Reddy
Parameter-efficient fine-tuning recipe for low-resource Indic translation with < 2% of the data required by baseline approaches.
EMNLP 2023 WS
PDF soon

Work with our research team.

We partner with universities, research labs and R&D departments on applied problems. If you have a real-world system and a research question attached, let's talk.

Track 01

Joint publications

Co-author applied-research papers with our founders and published scientists in your org.

Track 02

Thesis partnerships

We host 1–2 masters candidates per year on funded research projects tied to our live systems.

Track 03

R&D consulting

A 4–8 week applied research engagement for enterprises with a specific deep-tech question.

Have a research
problem we'd love?

contact@firstpointstudios.comReply from a founder within 48h
Start a conversation