Overview
News
Technologies
Salaries
Products
People
Growth
Offices
Jobs
Financials
Funding and backers
Team
11-50
employees, estimated range
Hiring now
2
+1 posted in 30 days
Momentum
13/100
hiring for its size
Last sign of life
this month
a posting or a news item
The story so fareach stage dated by its source: the company's releases, the press, registers, filings
May 29
On the Hacker News front page Real-time LLM Inference on Standard GPUs: 3k tokens/s per request · 219 points
2023
Founded
Tractionpublic signals, read monthly
Web presence
top 7%
of 133.2M domains by links, Common Crawl
Hacker News stories
5
best: 219 points
Investors in similar startupsfunds that back the companies most like this one
Similar startupsby what they do, their industry and backers
Makora inference endpoints use AI agents to optimize every layer of the stack — from GPU kernels and inference engines to algorithms and…
Hyde is the training and inference stack that allows enterprises to build their own specialist models.
The unified AI inference stack - from custom GPU kernels to production cloud serving on NVIDIA and AMD. 2x performance. Top open models…
Inferact is a startup founded by creators and core maintainers of vLLM, the most popular open-source LLM inference engine. Our mission is…
Open source inference engine for the hardware you already own. Apple Silicon, NVIDIA, AMD, or nothing but a CPU.
GPU orchestration, model inference, and an agent platform. Adopt Paladin, Ion, or Talos independently, or connect the full Cumulus stack.
Overview
Kog Labs builds a co-designed inference stack for real-time AI agents on standard datacenter GPUs, spanning model architecture, inference engine, compilers and low-level GPU kernels. It develops the Laneformer 2B model and states a 30 times faster LLM inference through parallel architecture. The company is headquartered in Paris, France.
News
Kog Laneformer 2B: The Latency-First Model Behind Kog Inference Engine
Today Kog is releasing the weights and model code of Laneformer 2B on Hugging Face Hub, the 2.3B-parameter instruction-tuned coding model designed for high-speed decoding.Most LLM research optimizes for benchmark quality first, and inference metrics like
Read more
Report
Real-time LLM Inference on Standard Datacenter GPUs (3,000 tokens/s per request)
Today, Kog AI launches a tech preview of the Kog Inference Engine (KIE): 3,000 output tokens/s per request on 8 AMD MI300X GPUs and 2,100 on 8 NVIDIA H200 (FP16, no speculative decoding). This preview runs a 2B model, with support for large third-party Mo
Read more
Report
Delayed Tensor Parallelism for Faster Transformer Inference
DTP is a new Transformer architecture that hides communication overhead behind computation and weight streaming, enabling significantly faster batch-size-one inference on AMD and NVIDIA GPUs.
Read more
Report
Kog Reaches 3.5 Breakthrough Inference Speed on AMD Instinct MI300X
Kog AI breaks conventions with a radical approach to inference, free from legacy constraints, to address real market pain points
Read more
Report
Tech stack
Technologies named in job postings at Kog Labs over the last 12 months, by layer, with the share of postings for each.
Technologies in use
13
Postings analysed
3
Stack modernity
72/100AI adopter
Languages
Sass
Frameworks & libraries
CUDA
Transformers
MLIR
PyTorch
AI & ML
LLM
DeepSeek
Sign in to see how your skills match this stack
Sign In
Full Kog Labs tech stack
13 technologies · by team · pay by technology · changes · similar stacks
Growth
Hiring Momentum
77/100
Scaling fast
Open positions
2
1 opened / 0 closed in 30 days
Median time-to-fill
110 days
faster than 12% of the market
ATS activity
Every ~4 hours
Today
Hiring Dynamics
+100%
Hiring Focus
The percentage next to each role is its share of the company's job openings over the last 90 days; the arrow shows the shift versus the previous period.
AI/ML
100% ▲
Activity Timeline
Added MLIR to stack
Sep 2026
Added Triton to stack
Sep 2026
Added Sass to stack
Sep 2026
Added HPC to stack
Sep 2026
Offices
Where the company hires and what each office is for
Cities
1
Hiring now
1
Countries
1
Busiest office
Paris
Paris
France
HQ
2 open roles 1 posted in 90 days 100% engineering 10-50 est. staff
hires for AI/ML
Jobs
≈ $57k – $123k per year (Estimated) • Hybrid • Full-Time • PhD • Paris
Quantization
AI Agents
Transformers
LLM
CUDA
MLIR
Speculative Decoding
Apply
≈ $57k – $122k per year (Estimated) • Hybrid • Full-Time • PhD • Paris
Quantization
AI Agents
Transformers
LLM
Mixture of Experts
Speculative Decoding
Apply