PUBLISHED AI RESEARCH AND PROJECTS

AI research and projects

The research, projects, and explanations behind the people. Read the source, inspect verified code and models, and find a collaborator.

Clear all filters

A source check confirms that a source supports the summary. It is not an independent replication or endorsement. Our methodology ↗

Showing 1–16 of 241 proofs of work · Kind: implementation · Newest first
Page 1 of 16
implementationChecked Sep 25, 2026

Accenture Foundation Model Adoption & Value Realization Framework

Lan Guan · Accenture

Developed Accenture's proprietary enterprise AI reinvention methodology, deploying foundation models across supply chain, customer service, and software engineering for Fortune 500 clients while benchmarking return on AI investment (ROAI).

Scope & limitations

Large-scale enterprise legacy IT modernization timelines often outlast fast-moving generative foundation model upgrade cycles.

Context: Enterprise LLM operationalization, multi-cloud foundation model architectures, ROAI metrics.

View mission
implementationChecked Sep 25, 2026

Databricks Mosaic AI & Lakehouse Architecture for Foundation Models

Ali Ghodsi · Databricks

Led the engineering integration of MosaicML training and fine-tuning pipelines directly into the Databricks Lakehouse, enabling organizations to pretrain and align custom open-weight models securely on internal enterprise data lakes.

Scope & limitations

On-premise and private cloud cluster provisioning constraints can bottleneck distributed tensor parallelism across non-homogeneous GPU topologies.

Context: DBRX, MPT-7B/30B, MosaicML Composer, Delta Lake, Vector Search.

View mission
implementationChecked Sep 25, 2026

Deterministic Agentic Execution Architecture for Enterprise Customer Systems

Bret Taylor · Sierra / OpenAI

Architected Sierra's Agent OS, combining generative LLM natural-language comprehension with deterministic state machines and strict transactional API sandboxes to prevent agent hallucination in consumer-facing enterprise tasks.

Scope & limitations

Enforcing strict deterministic guardrails constrains open-ended agent creativity and requires substantial domain-specific schema modeling.

Context: Multi-LLM routing, deterministic state automata, transactional CRM/ERP webhooks.

View mission
implementationChecked Sep 25, 2026

Enterprise Knowledge Graph Grounding for Multi-Source RAG Systems

Arvind Jain · Glean

Architected Glean's enterprise intelligence engine, coupling deep permissions-aware crawling of 100+ SaaS applications with dense vector embeddings and a dynamic enterprise knowledge graph for sub-second grounded generative question answering.

Scope & limitations

Maintaining real-time permission sync across complex Active Directory, Okta, and SaaS permission structures introduces significant distributed indexing overhead.

Context: Hybrid sparse-dense retrieval, enterprise permissions ACL enforcement, LLM synthesis.

View mission
implementationChecked Sep 25, 2026

Lamini Memory Tuning: Precise Fact Recall Without Hallucination

Sharon Zhou · Lamini / ex-Stanford

Developed Lamini Memory Tuning, a technique embedding millions of precise factual enterprise associations into LLM weights with zero hallucination while preserving generalized instruction-following capability.

Scope & limitations

Aggressive factual weight baking risks model overfitting and can degrade out-of-distribution reasoning on non-memorized prompts.

Context: LoRA/PEFT adaptation, cross-entropy weight tuning, hallucination evaluation harnesses.

View mission
implementationChecked Sep 25, 2026

NVIDIA Blackwell Platform: Accelerated Architecture for Trillion-Parameter Generative AI

Jensen Huang · NVIDIA

Architected the Blackwell B200 GPU and NVLink 5 switch chip infrastructure, introducing second-generation Transformer Engine FP4/FP8 micro-tensor scaling and high-density liquid-cooled rack computing designed to reduce trillion-parameter LLM inference costs and energy consumption by up to 25x.

Scope & limitations

High thermal design power (up to 120kW per NVL72 rack) requires specialized data center liquid-cooling retrofits and high-density power distribution.

Context: NVIDIA Blackwell B200, NVLink 5 (1.8TB/s bidirectional interconnect), FP4/FP8 Transformer Engine, TensorRT-LLM.

View mission
implementationChecked Sep 25, 2026

OpenAI o1 & Test-Time Compute Scaling Architecture

Sam Altman · OpenAI

Directed the commercialization and architectural rollout of OpenAI o1, shifting foundation model capability scaling from purely pretraining tokens to inference-time chain-of-thought reinforcement learning reasoning loops.

Scope & limitations

Increased inference latency (tens of seconds per complex query) and higher token costs make pure test-time reasoning impractical for real-time conversational streaming without hybrid routing.

Context: OpenAI o1, chain-of-thought reasoning, reinforcement learning from human and rule-based feedback (RLHF/RLAIF).

View mission
implementationChecked Sep 25, 2026

RedPajama: Open Pretraining Datasets for Reproducible Foundation Models

Vipul Ved Prakash · Together AI

Co-led the open-source release of RedPajama, a 1.2-trillion and 30-trillion token pretraining dataset replicating the LLaMA pretraining corpus to enable transparent, fully reproducible foundation model research.

Scope & limitations

Web-scale data deduplication and filtering at 30T scale require petabyte-scale distributed compute and risk residual benchmark contamination.

Context: 30T token web corpus, deduplication pipelines, MinHash LSH, toxicity filtering.

View mission
implementationChecked Sep 25, 2026

Salesforce Agentforce: Autonomous Enterprise Multi-Agent Orchestration

Clara Shih · Meta / ex-Salesforce

Spearheaded the product architecture of Agentforce, enabling enterprises to deploy autonomous AI agents with unified CRM data grounding, Atlas reasoning engines, and policy guardrails that operate without rigid hardcoded decision trees.

Scope & limitations

Enterprise agent reliability requires high data cleanliness and granular field-level permissioning across fragmented legacy enterprise systems.

Context: Atlas Reasoning Engine, Einstein Trust Layer, CRM Data Cloud grounding.

View mission
implementationChecked Sep 25, 2026

Snorkel: Rapid Training Data Creation with Weak Supervision

Alex Ratner · Snorkel AI / University of Washington

Created the Snorkel open-source system and paradigm, using generative probabilistic modeling to combine noisy heuristic rules, knowledge bases, and small models into high-accuracy probabilistic training labels without manual annotation.

Scope & limitations

Heuristic labeling functions can introduce correlated bias if domain experts write overlapping or conditionally dependent heuristics without proper graph independence modeling.

Context: Probabilistic generative label models, matrix completion, weak supervision algorithms.

View mission
implementationChecked Sep 25, 2026

Together Inference Engine & FlashAttention-Optimized Model Serving

Vipul Ved Prakash · Together AI

Engineered Together AI's custom inference kernel stack, implementing speculative decoding, custom FP8 FlashAttention kernels, and continuous batching to deliver the industry's fastest serving latency for Llama 3, DeepSeek, and Mixtral.

Scope & limitations

Highly optimized custom kernels require aggressive quantization and continuous hardware profiling for non-standard GPU architectures.

Context: Llama 3/3.1, Mixtral 8x22B, custom CUDA/Triton kernels, speculative sampling, continuous batching.

View mission
implementationChecked Sep 24, 2026

Agent.ai: Multi-Agent Professional Workflow Directory & Execution Engine

Dharmesh Shah · HubSpot / Agent.ai

Architected a global discovery and execution platform for specialized AI agents that orchestrate multi-step research, prospect intelligence, competitive synthesis, and automated business workflows.

Scope & limitations

Autonomous agent execution requires strict human-in-the-loop review for state-changing enterprise actions to prevent cascading hallucinated writes.

Context: Distributed agentic orchestrator, LLM tool-calling APIs, dynamic web sandboxes.

View mission
implementationChecked Sep 24, 2026

ChatSpot.ai: Natural Language Conversational CRM Layer

Dharmesh Shah · HubSpot / Agent.ai

Developed one of the first production enterprise generative AI integrations combining OpenAI GPT-4 with structured enterprise CRM relational databases, executing natural language SQL queries and automated prospecting.

Scope & limitations

Schema complexity across enterprise custom objects necessitates strict few-shot prompt guards to avoid incorrect relational joins.

Context: GPT-4, HubSpot API, semantic intent parsers, relational data synthesizers.

View mission
implementationChecked Sep 24, 2026

Diffusers: State-of-the-Art Diffusion Models in PyTorch

Suraj Patil · Hugging Face

Engineered core schedulers (DDIM, DPMSolver, Euler), memory-efficient cross-attention kernels, and quantized pipeline architectures powering Stable Diffusion and Flux ecosystem inference.

Scope & limitations

Diffusion sampling requires multiple sequential denoising steps, bottlenecking latency relative to single-pass adversarial models.

Context: PyTorch, CUDA, xFormers, FlashAttention, ONNX runtime.

View mission
implementationChecked Sep 24, 2026

Diffusion Model Distillation & Latent Consistency Optimization

Sayak Paul · Hugging Face

Co-authored and integrated Latent Consistency Model (LCM) LoRA adapters into Hugging Face Diffusers, compressing multi-step diffusion sampling into 2 to 4 inference steps for real-time generation.

Scope & limitations

Aggressive 2-step distillation can introduce subtle high-frequency smoothing in photorealistic background details.

Context: PyTorch, CUDA, LCM-LoRA, SDXL, Apple Silicon Metal.

View mission
implementationChecked Sep 24, 2026

GitHub Copilot: Real-Time Inline Generative Code Completion Architecture

Alex Graveley · Minion AI / Former GitHub

Conceived, architected, and built the initial GitHub Copilot production system, engineering client-side AST context window chunking, token speculative decoding, and low-latency streaming to VS Code.

Scope & limitations

Context window constraints in early versions restricted completions to single-file buffers before repository-wide semantic indexers were introduced.

Context: OpenAI Codex, TypeScript, VS Code language client, telemetry evaluation loops.

View mission
Showing 1–16 of 241 proofs of work