# Mem0 Raises $24M and Expands San Francisco Team to Build AI Agent Memory

**Author:** Zero G Talent
**Published:** 2026-11-15
**Category:** engineering
**Tags:** yc-hiring, developer tools, infrastructure, software engineering, generative ai, machine learning

[Read on Zero G Talent](https://zerogtalent.com/blog/mem0-raises-24m-and-expands-san-francisco-team-to-build-ai-agent-memory)

---

Most large language models operate statelessly, forgetting previous context once a session ends or context windows overflow. [Mem0](https://zerogtalent.com/ai-companies/mem0) develops an external memory layer designed to let AI agents retain, update, and retrieve facts and preferences across interactions. The San Francisco company raised $24 million in combined Seed and Series A financing to expand its infrastructure and scale its core engineering group.

The financing was led by Basis Set Ventures, alongside Peak XV Partners, Kindred Ventures, GitHub Fund, and Y Combinator, with backing from executives at Datadog, Supabase, PostHog, and Weights & Biases, as reported by [TechCrunch](https://techcrunch.com/2025/10/28/mem0-raises-24m-from-yc-peak-xv-and-basis-set-to-build-the-memory-layer-for-ai-apps/). To support enterprise adoption and research, Mem0 is hiring onsite engineers across four open positions in San Francisco, including a [Research Engineer - Agent Memory](https://zerogtalent.com/ai-jobs/mem0/research-engineer-agent-memory-99945) and a [Forward Deployed Engineer](https://zerogtalent.com/ai-jobs/mem0/forward-deployed-engineer-99951).

## The Architecture of Persistent Agent Memory

Mem0 provides both an open-source framework and a managed cloud platform that integrates into AI applications using a few lines of code. Instead of passing entire conversation transcripts into every model prompt—which drives up token costs and latency—the platform extracts high-signal facts and stores them across a hybrid data setup. As detailed by [InfoWorld](https://www.infoworld.com/article/4026560/mem0-an-open-source-memory-layer-for-llm-applications-and-ai-agents.html), the framework coordinates vector databases for semantic retrieval, key-value stores for direct factual lookups, and graph stores for relational mapping.

In a presentation for [Arize AI](https://www.youtube.com/watch?v=EE4pvOEAjXc), co-founder and CEO [Taranjeet Singh](https://zerogtalent.com/people/taranjeet-singh) explained that Mem0 runs two deterministic pipelines: an extraction pipeline that isolates memories from message pairs and past summaries, and an update pipeline that modifies or deletes stale references. On the cloud platform, Mem0 scores memories by relevance, importance, and recency to determine what should be retrieved or deprioritized. 

Beyond flat factual storage, the team published research on [alphaXiv](https://www.alphaxiv.org/abs/2504.19413) introducing Mem0g, an extension that organizes entities and relationships into directed graphs. In evaluations on the LOCOMO benchmark, the base Mem0 architecture showed a 91% reduction in p95 latency and more than 90% savings in token usage compared to full-context methods, alongside a 26% improvement over OpenAI's memory system on LLM-as-a-Judge evaluations.

## Technical Leadership from Tesla and Developer Tools

Mem0 was founded in 2023 by [Taranjeet Singh](https://zerogtalent.com/people/taranjeet-singh) and [Deshraj Yadav](https://zerogtalent.com/people/deshraj-yadav), who have collaborated on open-source projects for more than a decade. The duo previously built EvalAI, an open-source machine learning benchmark platform that began during Yadav's master's research at the Georgia Institute of Technology.

Singh started his engineering career at Paytm before becoming the first growth engineer at Khatabook. In late 2022, he created Cookup AI, an early GPT app directory that reached more than one million users, followed by Embedchain, a retrieval framework that accumulated thousands of GitHub stars. In an interview on the [InProd](https://www.youtube.com/watch?v=ljb7zd2nzfk) podcast, Singh noted that user feedback on an AI meditation application built on Embedchain directly sparked Mem0: users complained that the assistant could not recall their ongoing habits and context.

Yadav serves as chief technology officer. Before Mem0, he led the AI Platform at Tesla Autopilot, where he managed large-scale training pipelines, model evaluation, and observability infrastructure for full self-driving development. He has published research in conferences including CVPR, ECCV, and AAAI.

## Ecosystem Integrations and Enterprise Deployment

Mem0 has accumulated more than 41,000 GitHub stars and 14 million downloads. Cloud API usage grew from 35 million requests in the first quarter of 2025 to 186 million requests in the third quarter, according to [BW Disrupt](https://www.bwdisrupt.com/article/mem0-raises-24-mn-from-basis-set-ventures-peak-xv-partners-others-577381).

Cloud providers and agent frameworks have adopted the platform directly. AWS selected Mem0 as the memory provider for its Agent SDK. In technical documentation from [Amazon Web Services (AWS)](https://aws.amazon.com/blogs/machine-learning/company-wise-memory-in-amazon-bedrock-with-amazon-neptune-and-mem0/), cybersecurity company Trend Micro demonstrated using Mem0 alongside Amazon Bedrock and Amazon Neptune to equip its enterprise support chatbot with persistent, organization-specific context. Another architecture outlined by [Amazon Web Services (AWS)](https://aws.amazon.com/blogs/database/build-persistent-memory-for-agentic-ai-applications-with-mem0-open-source-amazon-elasticache-for-valkey-and-amazon-neptune-analytics/) showed Mem0 orchestrating vector search in Amazon ElastiCache for Valkey and relational graphs in Amazon Neptune Analytics to drop agent research query tokens twelvefold.

Independent tutorials by [MarkTechPost](https://www.marktechpost.com/2026/04/15/how-to-build-a-universal-long-term-memory-layer-for-ai-agents-using-mem0-and-openai/) highlight Mem0's multi-user isolation features, illustrating how developers scope distinct memory namespaces to preserve privacy and structure in multi-tenant environments.

## Engineering Roles and Onsite Collaboration

Mem0 operates as a compact four-person team based in San Francisco, with four open full-time technical roles to support platform scaling and client implementations. All current openings require onsite work in San Francisco.

On the model and system side, the [Research Engineer - Agent Memory](https://zerogtalent.com/ai-jobs/mem0/research-engineer-agent-memory-99945) reports directly to CTO [Deshraj Yadav](https://zerogtalent.com/people/deshraj-yadav). The role owns memory features end-to-end, focusing on fine-tuning models for extraction, updates, and consolidation while moving research concepts into production environments.

To bridge research and customer delivery, the [Forward Deployed Engineer](https://zerogtalent.com/ai-jobs/mem0/forward-deployed-engineer-99951) embeds with strategic customer accounts to design, build, and deploy memory API integrations into production AI agents. This role includes occasional travel for customer engagements.

Platform development is supported by the [Full Stack Engineer](https://zerogtalent.com/ai-jobs/mem0/full-stack-engineer-97144), who also reports to Yadav and owns platform features across Next.js and Python from design through deployment. Community and developer growth are managed by the [Developer Relations](https://zerogtalent.com/ai-jobs/mem0/developer-relations-99947) role, responsible for open-source documentation, technical integration guides, and conference representation.

## Open Roles at Mem0

| Role | Location | Work Mode | Salary |
| --- | --- | --- | --- |
| [Research Engineer - Agent Memory](https://zerogtalent.com/ai-jobs/mem0/research-engineer-agent-memory-99945) | San Francisco, CA, US | Onsite | $175,000 – $250,000 |
| [Full Stack Engineer](https://zerogtalent.com/ai-jobs/mem0/full-stack-engineer-97144) | San Francisco Bay Area | Onsite | $150,000 – $250,000 |
| [Forward Deployed Engineer](https://zerogtalent.com/ai-jobs/mem0/forward-deployed-engineer-99951) | San Francisco, CA, US | Onsite | $150,000 – $190,000 |
| [Developer Relations](https://zerogtalent.com/ai-jobs/mem0/developer-relations-99947) | San Francisco, CA, US | Onsite | $140,000 – $180,000 |

[See all open roles at Mem0](https://zerogtalent.com/ai-companies/mem0)

---
*Source: [Zero G Talent](https://zerogtalent.com), the job board for space, defense, AI, robotics and frontier-tech companies.*