TROOP

Software Engineer

NextRetreat

- Conducted regular customer and audience discovery calls to gather and synthesise information to build opportunities and test assumptions to craft the roadmap.
- Increased monthly active users by 400% after changing the approach.
- Major shift in approach: customer obsession and product-led-growth (PLG).
- Implemented multiple recommendation algorithms, machine learning models into the platform
- Shipped multiple products from idea to productiom
- Utilised geospatial knowledge to build a robust recommendations based on geography data.
- Designed and executed a new, container-based architecture utilizing Terraform, migrated from render.com to AWS and built on ECR and ECS, ensuring seamless operation and scalability.
- Established a robust CI/CD pipeline that supported testing and deployment of interdependent projects with GitHub Actions, streamlining development and release processes.
- Developed a serverless URL shortener using AWS API Gateway and DynamoDB, delivering an efficient and cost-effective solution.
- Managed multiple background jobs processing data from external sources, ensuring the integrity and reliability of the data.
- Created a framework to efficiently deploy long-running tasks to ECS using Terraform, optimizing system performance and scalability.
- Implemented optimization measures on the architecture and database, resulting in a significant reduction of costs by 35% and stable utilization of CPU/RAM resources.

Chief Technological Officer

Softygon

* Acquired by HUGLO Europe *
- Developed a universal platform for seamless software integration, leveraging cutting-edge technologies to facilitate interoperability and data exchange.
- Designed and deployed a serverless architecture utilizing AWS Lambda (Node.js), AWS RDS PostgreSQL, AWS SQS, AWS API Gateway, and AWS S3, creating a scalable solution for connecting APIs.
- Achieved seamless integration with hundreds of APIs, enhancing the platform's functionality and performance.
- Delivered a highly scalable and resilient architecture capable of processing large volumes of API requests, ensuring the platform's stability and reliability.

Chief Executive Officer

Chatbotech

- Designed and built a next-generation platform for constructing chatbots with integrated artificial intelligence, delivering advanced capabilities and enhanced user experience.
- Executed the implementation of new features in the Node.js back-end and Angular 2+ front-end, optimizing system performance and scalability.
- Facilitated seamless integration with external APIs, streamlining data exchange and augmenting the platform's capabilities.
- Demonstrated expertise in Kubernetes, providing hands-on experience in managing large-scale, highly available and resilient applications.

Full-stack Developer

Profit365

- Administered the MS SQL Server database on an on-demand EC2 instance, ensuring seamless operation and data security.
- Expertly maintained and optimized the infrastructure on EC2 instances, leveraging cloud computing technologies to deliver high availability and scalability.
- Implemented strategic optimization measures on the database, resulting in a significant reduction of costs by 40%, enhancing overall efficiency and productivity.

Lead Developer

- Designed and executed complex software solutions, integrating new functionalities into the back-end using .NET and enhancing the front-end user experience through AngularJS and web forms.
- Led the development and maintenance of a robust subscription system, ensuring its stability and scalability.
- Developed a cutting-edge Windows desktop application to streamline point-of-sale operations.
- Designed and deployed a cross-platform mobile application, leveraging cutting-edge technology to deliver a seamless user experience.
- Created a cutting-edge Apple Watch application utilizing Swift programming language to provide enhanced functionality and an intuitive user interface.

CPTO @ NextRetreat | Specializing in SaaS Growth and Promoting Sustainable Travel

I am a highly motivated individual with a background in Artificial Intelligence. Throughout my career, I have worked on multiple SaaS projects, including my most recent role as CPTO at NextRetreat, a retreat planning tool.

At NextRetreat, I have focused on driving growth and user activation through the implementation of various frameworks. In addition to my technical expertise, I am also passionate about promoting sustainable travel and have made it a focus of our company.

With a strong background in applied informatics and a track record of success in the SaaS industry, I am excited to continue growing and learning in my career.

Peter Kulcsár Szabó

CEO at Slite (YC18) | We build the first knowledge base on autopilot, in remote since 2016!

I'm founder and CEO at Slite (YC W18) & Super 👉 Super.work 

Super.work turn all your company data into answers & AI agents for your whole team.

It comes with the recipes and tools for your team to get instant benefit from AI at work

Slite is your team dead simple yet crazy powerful knowledge base, since 2017 🔥

I'm fortunate to lead a diverse team of exceptional individuals spread across 10+ countries 🇮🇳 🇬🇷🇱🇹 🇦🇷 🇬🇧 🇳🇱 🇫🇷 🇳🇴 🇺🇸 🇪🇸 🇩🇪 🇨🇦 🇵🇹 . Our journey includes going through YC W18 batch, raising $15M+ in funding from top-tier investors like Spark Capital and Index Ventures, and serving tens of thousands of teams monthly.

Christophe Pasquier

Super 2.0 is live, and it’s the release I’m the most proud of 🚨🚨
V1 launched just six months ago and found product-market fit fast. We won head-to-head deals against much bigger competitors.

And yet, we've done a complete rebuild.

Now, Super understands your question's nuances like a colleague would.
Super thinks, plans, and calls multiple tools intelligently for every answer. 

It became a true orchestrated search, able to combine semantic, structured search, web queries, iterative steps, templated outputs, grep, chart generation.

It gives the depth level of deep research but on your team data, and in a RELIABLE way. And it's still 40% faster than V1.
It's also 2x more accurate than our closest competitor, 2.3x more accurate than V1.

And all these improvements unlocked the feature I'm arguably the most excited about - Workflows.

I've deliberated on calling it "knowledge workflows" or "pre-set orchestrated searches" but my marketing team hates me when I throw these and there's no common term for "Workflows for people who don't have the time to mess around with workflow builders, webhooks, and nightmarish APIs."

But well, workflows do just that. Explain your needs and steps in 2 sentences ("Grab my help center article and compare it to our Intercom tickets to see what's missing", "Build a full customer overview from Attio, Slack and Slite") and Super will build an assistant, ready to generate reports on demand, or to generate perfect answers, in the right format, on the right data for your team.

I'm excited about Workflows, especially, because it actually enables companies become AI-native without having everyone to be on top of Zapier, n8n, configuring webhooks, and whatnot.

V2 doesn't force you to train your workforce or force them to change how they work. People just get their work done faster.
Okay. Rant over.

All in all,
I'm so freaking  proud of what the team has built. 
This is the best company search in the world. 
The most accurate, the fastest, and the easiest to deploy to solve true business problems on day 1. 
Built by an incredibly mighty team of 20 people, across 10 countries. 

Aaaaaand we're live on Product Hunt. So if you have a minute, go support the team and let's talk to bring Super in your company.

CEO Slite & Super (W18) | Most accurate search & workflows on team data

Anthropic is going after Microsoft.
OpenAI is going after EVERYONE.
In the last few months has OpenAI has
1. Hired an army of ex-Meta heads to work on monetisation and (soon) show you personalised ads within ChatGPT responses
2. Acquired Windsurf to compete with Copilot
3. Acquired Sky to control your Mac
4 Launched ChatGPT Atlas
5. Made Sora 2 as an entry-point to social short-form content

In the meantime,
Anthropic
1. Has surpassed OpenAI in enterprise usage (32% API usage vs OpenAI's 25%)
2. Drastically improved Claude Code, miles ahead MS' Github Copilot
3. Launched doc/spreadsheet/slide generation right within Claude
4. Launched Claude in Excel

OpenAI going after everyone is obvious. 
But Anthropic's moves feel more... subtle. They're quietly capturing business workflows and might disrupt Microsoft's Enterprise monopoly.
Thoughts?

"RAG is dead" was the stupidest statement I heard in SF

I've heard this a lot in the last few months. And every time the person meant
"Embeddings + GenAI" is dead.

Many implemented naive versions of it in 2023/24 and were *surprised* when it failed.

Like agents don't mean Zapier + LLMs, 
RAG just means Retrieval, i.e. search + AI generation.

And while context windows explode (and many think "let's shove all our data there!")
1/ It won't fit. If you don't know where data lives, there is no way you can put all your team info in every single query you need to run.
2/ Also it would cost a lot
3/ Agentic workflows #1 bottle neck is autonomous access to accurate team info. Nothing will work without accurate and complete retrieval.
4/ Accuracy decrease with context size
All this to say, you need RAG.

And to break the assumption we listed above, Retrieval ≠ Embedding, or not only.

Retrieval can go far beyond this rudimentary conception and is taking EXTREMELY interesting shapes. There's
→ Claude code's way of 'progressive' retrieval
→ Browser agent's way of 'direct' retrieval
→ MCP's 'on-call' retrieval
→ SQL querying based retrieval
→ Ofc, the old school semantic and keyword search
→ And orchestrating all of the ☝️ to get even more complete results
Agentic workflows give slop (or time out) with bad retrieval.

They do wonders when connected to a smart data layer like Super.work.

With our latest version we're blending many retrieval techniques for different tasks to ensure Super gives the more accurate and complete answer you could get from any of the tools on the market.

Super v2 is the best f**king Agentic RAG implementation on the planet, and it's surely not dead 🙃

Wispr Flow is killing it so I wanted to find who designed it. Turns out - I did 😏
This is in 1 picture why AI fails your team. Tomorrow we fix it!

We'll be announcing our most ambitious release ever with Super v2.

Tommorrow I'll share what the fuss is about, but for now here is why it matters🔥

95% AI pilots fail, 50%+ Zapier/n8n workflows never make it past 2-3 runs, most AI SaaS don't make it past demos - and they all share the same pattern.

Week 1 → Demos look amazing. Week 2 → People try it on real work, works for the basics, breaks for the real applications. Week 3 → You're back to asking humans in Slack.

The problem isn't the AI or the tooling on top. It's the context layer.

Most AI tools and workflows get a very small part of your data. Either by manual uploads or MCPs and API calls, the context-feeding approach never scales to a point where AI knows enough about your company - and your preferences to do a good job.

- AI can summarise the last 10 deals, but not why your biggest leads this year dropped out by connecting insights from your CRM, sales recordings, and real Slack threads from your AEs.
- AI can edit a changelog entry, but not go into your Linear, product documentation, and Github PRs to make a comprehensive list of every bug fix, every new feature, and who it was requested by.
- More basic, AI can try to answer "How do I do X" by looking for these exact words, but it will fail at finding this answer in a  thread of comments, or in docs using even slightly different wordings.

You can keep trying to spoon feed context, but it's a fallacy to think a few docs will ever replicate the sheer knowledge and unstructured insights that go into every job from your teammate.
There's no human-brain equivalent for your company's work and workflows.
There's no shared, omni-channel context layer that can pull everything from everywhere based solely on the job to be done.

This is the context layer problem in AI work.
Tomorrow, we solve it completely.

PMF is the most talked-about concept in startups. It's also the most misunderstood.
I've been building for 7 years. Been through YC. Heard PMF explained by some of the best founders and investors in the world.

And you know what? Most explanations still leave you confused.
A few weeks ago at our offsite, Pierre (my CTO) did a 20-minute ELI5 on PMF for our entire team.

No fancy frameworks. No complex metrics. Just clear, simple thinking about what PMF actually means and how you know when you have it.

The whole room got it.

If you want to see how Pierre explained it, let me know in the comments. Happy to share the presentation 👇

taplio

The Battle

Peter Kulcsár Szabó

Christophe Pasquier

Followers

Posts

Engagements

Likes

Performance ScoreChristophe Pasquier's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Peter Kulcsár Szabó

Christophe Pasquier

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Christophe Pasquier's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreChristophe Pasquier's LinkedIn Performance