NewsCatcher
Turning the web into a structured database of real world events.
Software Engineer (Backend & LLM)
About the role
Functions:
- Design and implement new data pipelines tailored to a specific project or new product.
- Continuously improve data pipelines to make them more generic and modular, reducing the effort required for subsequent integrations.
- Write, optimise and refactor prompts and prompt-related pipelines.
- Conduct thorough testing and validation of data pipelines to ensure accuracy and consistency of data.
- Ensure pipelines are optimized for performance, scalability, and reusability to facilitate future projects.
Examples of day-to-day tasks
- We want to support people’s enrichment within the CatchAll Tool. Do research on what is available on the market that can help us enrich data, Double-check open databases, and verify whether there is already an existing solution available within the NewsCatcher’s code base. Plan and design the pipeline. Build a prototype and demo it to the whole team. Proceed to, test on Dev, write test cases around. Deploy on Prod, get feedback and improve.
- One of the users complained about a job being stuck. Debug the pipeline, try to understand where the problem comes from: promts, code, database. Find the issue and prepare a fix. Test it within the whole pipeline on Dev. Break some pipeline, fix and test again. Deploy on Prod, make clients happy.
- Prompt responsible for validating results is giving only 50% precision. Review the prompt, review the model. Do some prompt engineering. Realize that “with each new model I feel like LLMs become stupier”. Try another LLM provider, adapt the prompt. Test results, improve the accuracy and create a dataset to prove it. Get 70% precision, wait for compliments.
Experience:
- 3+ years of experience in B2B SaaS as Backend / Software / Data Engineer
- Strong systems thinker with attention to performance and scalability
- Comfortable working with both SQL and NoSQL databases
- Experience shipping LLM-based functionality into production
- Able to move from prototype → stable, maintainable architecture
Must Have
- Strong Python (including async workloads)
- Docker & Kubernetes
- RabbitMQ (or Kafka / PubSub / SNS / ActiveMQ)
- MongoDB / DynamoDB / Redis
- Hands-on experience with LLM frameworks
We Also Use
- Elasticsearch/OpenSearch
- PostgreSQL / MySQL
- GitLab
- AWS / GCP / DigitalOcean
- Jenkins
Nice to Have
- Experience building production AI agents
- Frontend experience (useful for product UI)
- Experience in API-first or DaaS companies
Compensation and Perks:
- Competitive salary and equity
- Up to 24 days of vacation & 16 days of sick leave/holidays (all fully paid)
- One meeting-free day per week
- Co-working Budget
- Training Budget
- We provide all the necessary equipment to work comfortably and efficiently from home.
- Yearly company retreats (2025 — Portugal, 2024 — Canary Islands, 2023 — French Alpes)
About the interview
Recruiter>Interview with CTO>Tech Interview
About NewsCatcher
Our mission
To help companies never miss important news
What is NewsCatcher’s niche?
News API for tech-savvy Enterprises that cannot afford to miss out on the news.
Ideal Customer Profile
Any company that has analysts who spend over 100 hours/ month finding insights in online news articles
Pain Points:
- Analysts spend way too much time searching for news and not anlyzing it
- Analysts are not scalable + human error
- Miss out on relevant information (false positives, noise, not enough time to go through everything)
Our Values
Data quality first
We’re data-as-a-service.
So, data quality should be our biggest priority and competitive advantage.
We fully own the data collection process: web crawling, scraping, data cleaning, and quality assurance.
We provide Service Level Agreements on source availability and coverage; We add new publicly available sources when clients ask.
Product that is fast to adapt to customers’ needs
Failure to adopt the solution is what pushes companies away from our competitors.
There is no one-fits-all data enrichment/LLM/ChatGPT prompt.
Insight: we serve companies with over ten different use cases. There are multiple clients within the same use case, and each still has a unique approach to how they want to solve their problem.
So instead of focusing on a “magic pill” that can solve everyone’s problem just OK, we take a unique approach to make sure our customers are really happy and get a high Return On Investment.
Enterprise-level customer service
As you can see from our home page, we work with quite a lot of big logos (the majority we can’t display due to NDAs). Many big companies trusted us when NewsCatcher was just “two guys in a shed,” so we really try to pay them off with great customer support and service quality.
We know our numbers, and we’re the only solution on the market that provides full transparency to customers on such metrics as:
- average latency from the article being published to being indexed
- alarms on websites not being crawled & indexed properly
- uptime not just for API but also for the data pipeline that updates the index
We aim for high-security standards. We’re in the process of getting SOC2 Type 2 in 2024.
Engineers talk to customers
Most of our customers love us because they can talk directly to engineers. We usually set up Slack/Teams channels where customers can get the answer in just a few minutes.
Potential customers should get their questions answered on the first call
Yeah, we’ll still want to talk before giving access to our data, but the client (usually a techie) should get all questions answered, no matter how technical they are.
Only focus on what’s important
We came up with the company name in about 10 minutes of thinking.
We made our logo in 5 minutes by downloading Figma for the first time in our life and just typing </newscatcher>
Our website was hosted on a super-simple website builder called UMSO for ~$250/year until we reached a 7-digit Annual Recurring Revenue.
Our logo, our name, our website: none of these really matter. What matters is happy customers. Customers don’t care how good your logo is if you solve their expensive problems.
So that’s what the company focus should be: solving customers’ problems.
