Open Finance@Columbia

MOF, MOT, FinGPT, FinRL | Open Finance @ Columbia University | Academic member of Linux Foundation, PyTorch, LF AI & Data, FinOS, CRAFT

[Arklex AI @ArklexAI]

6 months of manual testing. Replaced in 30 minutes.

Jun-shuo (Lance) Liu , a research engineer at Columbia University, was stuck in a cycle most AI agent developers know well - designing test cases by hand, reading through every conversation, writing bug reports, and starting over with every update.

He tried ArkSim. Here's what happened:

→ Test report time: 2–3 days → 30 minutes
→ Iteration cycle: 1–2 weeks → 1–2 days
→ Accuracy: 80% → 90% in one week

But the biggest unlock wasn't speed. It was visibility. ArkSim surfaced a tool selection bug he'd been living with for months. This type of bugs was invisible to manual review, caught in a single run.

He wrote up the full story:
https://t.co/GUDKCLPDhn

#AIAgent #AIEval #AgentTesting

[Xiaoyang Liu @Xiaoyan42555703]

Find your agent's errors before your real users do! https://t.co/gOwDDjzQsP

[Computational Engineering, Finance, and Science @BNHNA]

zkFinGPT: Zero-Knowledge Proofs for Financial Generative Pre-trained Transformers

Xiao-Yang Liu, Ningjie Li, Keyi Wang, Xiaoli Zhi, Weiqin Tong
https://t.co/E1WhdBAtCL [𝚌𝚜.𝙲𝙴 𝚌𝚜.𝙲𝚁]

[Computational Engineering, Finance, and Science @BNHNA]

FinRL Contests: Benchmarking Data-driven Financial Reinforcement Learning Agents. https://t.co/mc6Lk6Jzf6

[Open Finance@Columbia @Ai4Finance]

Plz use your best judgement before real-life trading!

[Alpha Trader X @AlphaTraderX_]

This Free AI Made a Trader $1.2 Million!!
That’s exactly what happened when a Reddit trader used FinGPT — a free, open-source AI that scans real-time news, earnings reports, and tweets — to turn $280K in a week… & $1.2 million in 3 months.
#FinGPT #AITrading #FreeAI #AlgoTrading

[Forbes Russia @ForbesRussia]

«Финуслуги» от Мосбиржи представил обновленную версию FinGPT — своего инвестиционного ИИ-помощника. Он может предоставлять инвесторам информацию о происходящем на российском рынке, транслировать биржевые котировки, объяснять финансовые термины и тонкости…
https://t.co/exdBEMBqPN

[Computational Engineering, Finance, and Science @BNHNA]

Multimodal Financial Foundation Models (MFFMs): Progress, Prospects, and Challenges. https://t.co/rX4K5l2B8Z

[Open Finance@Columbia @Ai4Finance]

The new license OpenMDW is launched: https://t.co/0dN5gGxnib

[Open Finance@Columbia @Ai4Finance]

Getting frustrated with "#Openwashing" behaviors in #LLMs and #GenAI? #CloseAI vs. #OpenAI? Check out the Model Openness Framework #MOF at #Linux Foundation: https://t.co/QL9Ht6YqNa

[Computational Engineering, Finance, and Science @BNHNA]

FinLoRA: Benchmarking LoRA Methods for Fine-Tuning LLMs on Financial Datasets. https://t.co/GFBcccWxZn

[Financial Technology @fin_tech]

Parallel Market Environments for FinRL Contests https://t.co/M9v6ubExhh

[Quant Science @quantscience_]

💡FinRL: A Python package for Financial Reinforcement Learning

Let's explore: 🧵

[Open Finance@Columbia @Ai4Finance]

@Naman_Arora7 dm plz

[Open Finance@Columbia @Ai4Finance]

FinRL Contest 2025 starts! Welcome to explore https://t.co/vegaTC6dcB

[Computational Engineering, Finance, and Science @BNHNA]

Parallel Market Environments for FinRL Contests. https://t.co/mc6Lk6Jzf6

[DeepSeek @deepseek_ai]

🚀 Day 5 of #OpenSourceWeek: 3FS, Thruster for All DeepSeek Data Access

Fire-Flyer File System (3FS) - a parallel file system that utilizes the full bandwidth of modern SSDs and RDMA networks.

⚡ 6.6 TiB/s aggregate read throughput in a 180-node cluster
⚡ 3.66 TiB/min throughput on GraySort benchmark in a 25-node cluster
⚡ 40+ GiB/s peak throughput per client node for KVCache lookup
🧬 Disaggregated architecture with strong consistency semantics
✅ Training data preprocessing, dataset loading, checkpoint saving/reloading, embedding vector search & KVCache lookups for inference in V3/R1

📥 3FS → https://t.co/JRZ2eYk1aK
⛲ Smallpond - data processing framework on 3FS → https://t.co/BoUA6YNjZK

[DeepSeek @deepseek_ai]

🚀 Day 4 of #OpenSourceWeek: Optimized Parallelism Strategies

✅ DualPipe - a bidirectional pipeline parallelism algorithm for computation-communication overlap in V3/R1 training.
🔗 https://t.co/NR3047Nnpo

✅ EPLB - an expert-parallel load balancer for V3/R1.
🔗 https://t.co/IilrQUfn49

📊 Analyze computation-communication overlap in V3/R1.
🔗 https://t.co/tY1WujA1Lw

[DeepSeek @deepseek_ai]

🚀 Day 3 of #OpenSourceWeek: DeepGEMM

Introducing DeepGEMM - an FP8 GEMM library that supports both dense and MoE GEMMs, powering V3/R1 training and inference.

⚡ Up to 1350+ FP8 TFLOPS on Hopper GPUs
✅ No heavy dependency, as clean as a tutorial
✅ Fully Just-In-Time compiled
✅ Core logic at ~300 lines - yet outperforms expert-tuned kernels across most matrix sizes
✅ Supports dense layout and two MoE layouts

🔗 GitHub: https://t.co/cxJ55w61pT

[DeepSeek @deepseek_ai]

🚀 Day 2 of #OpenSourceWeek: DeepEP

Excited to introduce DeepEP - the first open-source EP communication library for MoE model training and inference.

✅ Efficient and optimized all-to-all communication
✅ Both intranode and internode support with NVLink and RDMA
✅ High-throughput kernels for training and inference prefilling
✅ Low-latency kernels for inference decoding
✅ Native FP8 dispatch support
✅ Flexible GPU resource control for computation-communication overlapping

🔗 GitHub: https://t.co/Q6eRxZ9kgW

[DeepSeek @deepseek_ai]

🚀 Day 1 of #OpenSourceWeek: FlashMLA

Honored to share FlashMLA - our efficient MLA decoding kernel for Hopper GPUs, optimized for variable-length sequences and now in production.

✅ BF16 support
✅ Paged KV cache (block size 64)
⚡ 3000 GB/s memory-bound & 580 TFLOPS compute-bound on H800

🔗 Explore on GitHub: https://t.co/4JvJTn5HX2

[Open Finance@Columbia @Ai4Finance]

@ivebotunac Appreciate your success in PhD dissertation! Thanks for your kind words! Hope you enjoy the journey forward!

[Open Finance@Columbia @Ai4Finance]

Great job!

[Ive @ivebotunac]

Just open-sourced PrimoGPT – an automated trading system using Deep Reinforcement Learning (DRL) & LLMs!

Built as part of my PhD dissertation, inspired by FinRL & FinGPT. It features an RAG-based approach for generating trading signals.

🔗 Repo: https://t.co/kIt4INyzr3

It's still a work in progress, but please feel free to check it out and contact me if you have any questions!