Product Manager & Owner with 5+ years building enterprise AI tools at the intersection of UX and technical complexity. I turn powerful but opaque systems into products people actually trust and use daily.
I'm Manuel, a Product Manager & Owner based in Berlin. My background is unusual: I started in advertising and photography, moved into UX, and eventually found my home in product. Specifically, the hard kind. The kind where the technology is genuinely powerful but the people using it can't figure out how.
For four years at Lengoo (a Berlin-based AI startup, $34M raised, building enterprise custom language models), I owned the product that let enterprise clients configure, train, and refine their AI translation systems. When Lengoo closed in 2024, I moved to Cognigy, one of Europe's leading conversational AI platforms, where I bridge frontline customer insights with product priorities.
My edge is this: I understand what AI products are doing under the hood, and I understand why users give up on them. That gap is where I work. Currently open to PM / PO roles at product-led companies where AI is not a buzzword but a real product challenge.
4+ years building enterprise AI platforms, from custom LLM tooling to conversational AI. I know what "production-ready" actually means for AI.
I start with the problem, not the solution. Interviews, ticket analysis, and usage data before any backlog item gets written.
Design isn't decoration. It's how complexity becomes usable. I apply UX thinking to decisions, not just interfaces.
CSAT, latency, adoption, repeat rate. I define success criteria before building, not after. Every roadmap item has a measurable goal.
Transformed a fragmented, ad-hoc ML configuration workflow into a guided enterprise platform, lifting CSAT by 15% and driving 20% adoption.
Real-time data pipeline that cut MT quality feedback latency by 85%, with automated P1 triage across 4 engineering teams via Jira API.
Creative studio trusted by Adidas, Zalando, and Blinkist. 100% on-time delivery across 20+ clients, 60% repeat business.
Built the HALOS Console brand from zero: logo system, colour identity, and all marketing materials. Brand thinking applied end-to-end.
Open to PM and PO roles at product-led companies in Berlin and remotely particularly where the product space involves AI, B2B SaaS, or complex technical tools. If you're building something meaningful and need someone who can own it end-to-end, I'd love to hear from you.
contact@manubecerra.comI led the end-to-end product lifecycle for the PI Model Customization initiative, a core component of Lengoo's HALOS Console. The goal was transforming a manual, fragmented workflow into a streamlined internal tool that enterprise clients actually trusted.
The HALOS Console platform was a foundation, but the core workflow for customizing machine learning models for language services was deeply fragmented. Engineers and specialized teams were relying on a series of ad-hoc scripts and fragmented interfaces. This led to a lack of visibility, slow model deployment, and a high volume of internal support inquiries.
The business challenge was to democratize access to our advanced AI capabilities and solve a critical operational bottleneck.
I architected a comprehensive solution strategy that delivered immediate impact while building scalable foundations for future growth. Every strategic decision was anchored in rigorous user research, quantitative analysis, and business outcome optimization.
Orchestrated comprehensive stakeholder research across 20+ interviews with linguists, product managers, and customer success teams to map user psychology, behavioral patterns, and critical friction points.
Synthesized qualitative insights with quantitative product analytics to architect a data-driven user story framework, ensuring engineering resources focused on maximum business impact features.
Architected comprehensive workflow blueprints, deconstructing complex model customization processes into clear, actionable user journey maps from conception to deployment.
Created organizational alignment through visual product documentation that became the strategic north star, enabling rapid decision-making and clear trade-off communication across all stakeholders.
Drove aggressive product execution through advanced agile methodologies, strategically managing backlogs and ruthlessly prioritizing features based on adoption and satisfaction KPIs.
Revolutionized release velocity by implementing bi-weekly deployment cycles (from monthly), accelerating time-to-market while driving measurable user engagement and business outcomes.
I delivered a transformational internal platform that eliminated chaos and unlocked AI potential, translating complex technical requirements into intuitive, scalable product experiences.
Engineered an enterprise-grade LLM management interface with intuitive tabular design, empowering users with real-time model oversight and streamlined operational control.
Architected a precision-engineered translation management interface that puts users in complete control of cost optimization and performance tuning.
Built comprehensive analytics dashboard with advanced visualizations, delivering instant performance insights that democratize data-driven decision making.
HALOS had powerful AI capabilities, but the ML model customization workflow was critically broken. Engineers struggled with fragmented interfaces and ad-hoc scripts, creating massive operational inefficiencies. I identified this as a strategic opportunity to unlock AI potential and eliminate a company-wide productivity drain.
"Project HALOS represents strategic product leadership delivering exponential business value, transforming organizational AI capabilities while establishing scalable frameworks for continuous innovation. This initiative became the catalyst for company-wide AI democratization and operational excellence."
End-to-end product strategy for Project ECHO, focused on resolving critical systemic latency in our MT model iteration cycle. I architected a data pipeline connecting structured user feedback directly to the engineering backlog, transforming a days-long manual process into a real-time triage engine.
Critical systemic latency in the MT model iteration cycle, spanning four teams: Linguistics, Front-End, Data Engineering, and ML. The manual feedback process added days of delay between a linguist identifying an error and an engineer acting on it. A direct threat to product quality and AI model improvement velocity.
The friction was causing expert linguists to disengage from the feedback loop entirely, starving the AI models of the high-quality training signal they needed. Our mandate: eliminate the latency, automate triage, and add a gamification layer to restore and sustain linguist engagement.
Vision: Transform the MT feedback-to-fix cycle from an organizational bottleneck into a competitive advantage: near zero-latency iteration, always-on quality signal.
Three concrete deliverables:
Collaborated with Data Science to define a structured error taxonomy. Used Kibana and Datadog analysis to identify the top 5 error classes driving 80% of rework These became the initial governance schema.
Result: objective, unambiguous Definition of Done. No more backlog refinement debates about severity.
Prioritized the Data Contract (JSON schema + API endpoints) as the very first sprint deliverable. This single artefact unlocked parallel development across all four dependent teams simultaneously, dramatically reducing sequential dependency risk.
Authored detailed technical specifications for the API ingestion layer and the Jira API integration, serving as the single source of truth for the development team. Technical depth here directly built trust with ML Engineering and accelerated sprint planning confidence.
Project ECHO delivered fundamental improvements across the entire MT quality pipeline:
"Total ROI: Project ECHO delivered measurable business value across multiple dimensions: reduced operational costs, faster time-to-market, improved product quality, and enhanced team efficiency. All while establishing a scalable, data-driven foundation for continuous improvement."
How I co-built a creative studio trusted by Adidas, Zalando, and Blinkist. Every project on time, 60% repeat business, and nine years of learning what it really takes to run a product from zero.
Aneekaa Studio was the creative agency I co-founded in 2015 and ran for nearly a decade, delivering brand strategy, digital design, and web development for clients across Spain and Germany. We worked with brands like Adidas, Zalando, Blinkist, and Fotografiska, and built a reputation for transparency, quality, and precision delivery.
Running Aneekaa wasn't just a creative role. It was a full product and business leadership role. I defined objectives, scoped work, estimated effort, coordinated a core team of three plus freelancers, and maintained client relationships through every project phase.
"Building from zero taught me that strategy without execution is theory, and execution without strategy is noise. You need both at the same time."
Running a studio for nine years before moving into formal product roles gave me something most PMs don't have. I've felt the full weight of delivery. I know what it means when a deadline slips, when a scope creeps, when a client loses confidence. That experience shapes how I think about roadmaps, estimates, and stakeholder communication today. I treat product decisions with the same seriousness as a founder would, because I was one.
I built the HALOS Console brand from scratch: logo, colour system, and all marketing materials. Then I applied that visual language consistently across the product, website, and sales assets. Brand strategy and product thinking working as one.
HALOS Console was Lengoo's enterprise AI platform. Powerful technology with no visual identity. It needed a brand that could work across the product UI, marketing website, sales decks, and external communications, while feeling technically credible and trustworthy to enterprise clients.
I owned this end-to-end: from initial positioning through logo construction, colour system definition, and every applied marketing asset. No external agency involved, built entirely in-house from concept to production.
The primary brand expression: the "hi" icon mark paired with the HALOS Console wordmark. Designed to work at any scale, from product UI to marketing materials.
Every enterprise product needs to work on both light and dark surfaces. The logo was designed so both versions use the same icon. Only the wordmark colour changes, maintaining full legibility and consistency.
The "hi" icon works as a standalone mark for small-format applications: product UI, favicons, app icons, social media. Built on a geometric grid so it scales cleanly from 16px to any size.
The logo went through an extensive exploration phase, testing geometric construction systems, gradient directions, and letterform variations. The construction grid shows how every curve and proportion is derived from the same circle geometry. Rigour over intuition.
Three primary brand colours, each with a specific role. Cyan (#0abce9) for vibrancy and tech energy, navy (#3f4497) for trust and enterprise credibility, blue (#007fe4) as the connecting mid-tone. Together they form a system that works across UI, print, and digital surfaces.
Building a brand from scratch forces every decision to be intentional. Colour didn't just look nice. It had to work in a data table, a loading state, a marketing headline, and a small icon badge simultaneously. That constraint-driven thinking is exactly how I approach product design.
The same visual system I designed here became the foundation for the HALOS Console product UI: colour tokens, type hierarchy, icon style. Brand and product weren't separate workstreams. They were one.
"A brand isn't a logo. It's the set of decisions that makes every touchpoint feel like it came from the same place, from the favicon to the sales deck. I built that system for HALOS Console and then shipped the product on top of it."
