Provectus

Responsible AI in Action. Balancing Innovation and Privacy. We built Barley as a simple experiment. But today it helps our teams organize projects, answer questions, uncover information, and stay on top of tasks. Being helpful is not enough, though. An AI assistant must also be safe. And Barley was designed with privacy first. ✔️ No collection of personal data unrelated to work ✔️ No use of personal data for model training ✔️ No access to confidential sources unless explicitly enabled ✔️ Clear consent for recordings ✔️ Human review for automated actions In our latest article, we share what we learned. From defining personal and confidential data to building an internal AI Assistant Usage Guide that aligns legal, technical, and operational safeguards. If you are building or deploying AI internally, this is a practical read. 👉 Read the full article on Medium: https://lnkd.in/dpDzwhnC

🚀 GenAI / Python Engineering Internship (3 months) Provectus helps companies adopt AI to transform how they operate, compete, and create value. We are looking for highly motivated interns to join our Ж-shaped (Z-shaped) Engineering team. We don’t just write code; we build end-to-end solutions that solve real business problems. This internship is designed for those who want to move beyond “just Python” and become versatile engineers who leverage the latest AI tools to deliver impact. Program Details & Timeline 🔸 Format: Hybrid/Off-line regular mentor sessions, practical tasks on real-world projects. 🔸 Locations: Yerevan (Armenia), Novi Sad, Belgrade (Serbia). 🔸 Goal: This is a high-bar program. Our goal is to offer Full-time positions to the top performers. Key Dates 🔹 Program Start: April 6th 🔹 Hiring Decisions: July 6th What You’ll Work On Our current focus is on building AI Solutions for the Finservices sector (Insurance, Hedge Funds, etc.). You will be involved in: ✔️ Agentic Systems & RAG: Building sophisticated systems that can reason, analyze massive document sets, and automate complex workflows. ✔️ Data Migration & Processing: Handling large-scale data transformation tasks necessary for production-grade GenAI. ✔️ End-to-End Development: Building backends (FastAPI/Flask), integrating with LLM APIs (OpenAI, Anthropic, AWS Bedrock), and ensuring scalability. What We Expect ◾ Technical Foundation: Proficiency in Python, understanding of Docker, and Git. ◾ GenAI Literacy: Experience with LLM APIs and a solid understanding of RAG principles. ◾ Engineering Mentality: A desire to build “real things” rather than just playing with notebooks. ◾ Soft Skills: Ability to learn at a high pace, work independently, and clearly communicate progress and blockers. ◾ Language: Upper-Intermediate English. How To Get Started? 1️⃣ Send your application: fill in all the required fields and attach your CV; 2️⃣ Receive and complete the entry technical assessment task to give us an idea of your current knowledge. You should submit it within 3 days after the interview to proceed; 3️⃣ Pass the initial interview with the recruiter; 4️⃣ Pass the technical interview; 5️⃣ Receive an invitation to the program; 6️⃣ Work hard, be curious, show good results, and you’ll get a job at Provectus! Don’t let this opportunity slip through your fingers - apply now! 👉 https://lnkd.in/d_B8xJMZ

Provectus was honored to join Anthropic’s Executive Briefing on Enterprise Agents alongside other strategic partners. Anthropic’s rapid advancement in the enterprise space is demonstrating a strong impact on how organizations transform with AI. We’re inspired by how Claude continues to push the boundaries of enterprise AI, providing systems integrators, like Provectus, with a powerful foundation to deliver measurable, AI-driven business outcomes at scale. We’re excited to be part of this momentum and to help enterprises turn these innovations into real value. #Anthropic #Provectus #EnterpriseAI #Agents

AI is writing nearly a third of all new code. Yet overall engineering productivity has barely moved. In our latest article, Konstantin Makarychev, our Director of Engineering, dives deep into the “AI Productivity Paradox” and explains why measuring lines of code or acceptance rates misses the point. Faster typing does not equal faster delivery. The Provectus Metrics Framework can help measure real Dev ROI, focusing on three signals that actually reflect business impact: ✔️ Cost per Merged Feature. The true marginal cost of delivering value. ✔️ Token Velocity. A proxy for delegation depth in agentic workflows. ✔️ Autonomous Execution Rate. How effectively agents write, test, and self-correct before human review. The key message is: Stop optimizing for code volume. Start measuring delivered outcomes, cognitive delegation, and system health. Read the full article to see how we measure the “Claude Effect” in practice. https://lnkd.in/dDk6Mbns

Proud to see our team in the room for Anthropic’s Executive Briefing on Enterprise Agents. As an Anthropic partner, we’re ready to take these conversations straight into delivery. What’s becoming clear is that enterprise value won’t come from generic AI assistants. It will come from custom "specialist" agents that understand your tools, your workflows, and your guardrails. That matters most in regulated environments where trust, auditability, and integration decide whether anything ships.

At Provectus, we keep a simple tradition. From time to time, our teams gather in the office for small, informal get-togethers. It’s a chance to pause. Step away from daily tasks. Talk about things beyond projects and deadlines. Share stories. Laugh. Spend time together as people, not just colleagues. These moments may seem small. But they strengthen what matters most. Our community. Here are a few highlights from our teams in Serbia and Ukraine, continuing this tradition of connection and collaboration.

Proud moment for our team. Provectus joined a select group of early partners at Anthropic’s offices for a hands-on Claude Code workshop with the Applied AI team. Grateful for the early access, the open dialogue, and the opportunity to collaborate closely on what’s next.

Drilling teams have the expertise they need but they're also buried under persistent inefficiencies in Oil & Gas. High-stakes decisions under tight HSE constraints, unstructured reports that hide critical insights, and downtime that quietly erodes millions per campaign leave operations reacting instead of optimizing. AI is already changing that for forward-thinking operators: ✔️ Turning raw Daily Drilling Reports into actionable intelligence ✔️ Spotting developing issues before they become costly incidents ✔️ Enabling real-time diagnostics and predictive interventions ✔️ Surfacing hidden inefficiencies across rigs and basins ✔️ Driving proactive, fleet-wide performance gains The result? Measurable NPT reductions, safer operations, and stronger capital efficiency. Explore the practical path forward in our latest guide: https://lnkd.in/djNKcSjF

Provectus is hiring AI talent in Colombia! 🌎 Location: Colombia | Remote-first We're looking for ML Tech Leads and Senior ML Engineers to build production-grade ML systems, LLM solutions, and RAG pipelines. You’ll design scalable architectures, mentor engineers, and work hands-on with cutting-edge AI in the cloud. Ready to innovate? Apply here: ✔️ Senior AI/ML Engineer - https://lnkd.in/dQunhdSd ✔️ AI/ML Tech Lead - https://lnkd.in/dREU3Tp4 Know someone who’d be a great fit? Just tag them below.

The Provectus team has successfully completed two engaging interactions as part of our ongoing Public Speaking initiative, and the journey has been amazing so far! From building confidence and overcoming stage fear to practicing powerful delivery and clear communication, participants are already stepping out of their comfort zones and finding their voice. This is just the beginning. More growth, more confidence, and more impactful speakers ahead!