Mariusz Lesniewski

AI Product Manager and discovery-led product leader with 13+ years of experience turning ambiguous business, product and workflow problems into clear product direction, validated pilots and shipped systems. I help organisations identify where AI can credibly improve efficiency, growth, decision quality or workflow reliability, then shape those opportunities into MVPs with baselines, evaluation, provenance, observability and explicit kill criteria.

I currently lead Signal To System, a boutique AI product discovery and transformation practice. My approach is deliberately practical: start with real business problems and process constraints, map how the organisation actually works, select the simplest credible solution, and only build when there is a clear way to test whether the system creates value.

My product background comes from years of finding signal in noisy, uncertain environments: prototypes, playtests, metrics, stakeholder pivots, user feedback and product direction under pressure. I bring that same discipline to AI initiatives: discovery first, product direction second, implementation only after the problem, value logic and validation path are clear.

• AI product discovery for ambiguous business, workflow and decision problems. Pre-sales and scoping support grounded in real product value, not only technical feasibility.
• Stakeholder and team interviews, change management, workshop facilitation, workflow mapping, organisational world-model / ontology building, process analysis and goal alignment.
• AI opportunity assessment: identifying candidates with the strongest efficiency, growth, decision-quality or reliability potential.
• MVP and pilot design with baselines, business/system/operational/trust metrics, provenance, evaluation, observability and kill criteria.
• Delivery-aware product partnership: hands-on MVP/pilot delivery to date, enabling effective collaboration with delivery managers and engineers around business value, data readiness, user needs, AI feasibility, delivery constraints and adoption risk.
• Human-centred AI: systems that augment accountable people rather than replacing consequential judgment.

1. Signal To System

   Founder / AI Product Discovery Lead & Consultant

   Warsaw, Poland · July 2025 – Present

   Boutique AI product discovery and transformation practice focused on helping organisations move from AI noise to practical systems that improve decisions, workflows, efficiency and growth.

   • Run AI discovery as a structured initiative design process: interviews, workflow mapping, data/source review, process analysis, goal alignment and opportunity prioritisation.
   • Build the organisation’s working model for each opportunity: actors, decisions, workflows, inputs, outputs, sources of truth, constraints, bottlenecks, ownership and feedback loops.
   • Prioritise AI candidates by problem reality, value potential, data readiness, evaluation clarity, technical feasibility, deployment realism and responsible AI fit.
   • Recommend the simplest credible solution class before reaching for LLMs or agents: process redesign, instrumentation, search/retrieval, rules, analytics, classical ML, bounded LLM workflows, deterministic AI workflows or agentic workflows where justified.
   • Shape Agile MVPs and pilots with baselines, good-enough thresholds, explicit kill criteria, human accountability and a path to test, improve, scale or stop.
   • Define provenance, evaluation and observability from the start: source grounding, versioning, logs, human feedback loops, outcome tracking and non-regression checks.

2. Selected AI discovery / applied AI portfolio cases

   • NSCLC evidence navigator, MIT AI/ML capstone, deployed MVP. Led most of the discovery, product direction, deployment and evaluation work for a high-stakes oncology evidence decision-support system. Designed a hybrid approach combining deterministic guardrails, semantic retrieval/RAG and LLM assistance, with provenance, citation checks, uncertainty, reviewer workflow and kill criteria. Focus: decision quality, risk reduction and workflow reliability, not clinical decision automation.
   • Leading Marketing & Events Agency, contract-to-payout workflow. Mapped a fragmented manual process across email, Excel, Drive, Asana, accounting and e-signature that forced outsourcing at ~30% additional contract/order cost. Rejected risky full LLM contract generation and recommended a deterministic in-house lifecycle with approved legal templates, status visibility, event trail, human review and operational control.
   • Confidential client, synthetic customer research / concept screening. Designed an AI-assisted preselection layer for early concepts and messaging before expensive panel research. Combined synthetic qualitative responses with Semantic Similarity Rating, rationale, uncertainty and provenance. Evaluation path: compare against real panel outcomes, stability across runs and correlation with later human benchmark results.
   • Major Hospitality & Entertainment client, traffic forecasting and staffing decision support. Built the discovery model for a multi-source operational forecasting problem across sales/visits, bookings, weather, events and staffing decisions. Recommended a 4-week MVP with retrospective validation against simple baselines before live use. Metrics: forecast error, under/overstaffed shifts, decision time, manager feedback and operational usefulness.

3. QLOC

   Product Director

   Warsaw, Poland · October 2023 – March 2025

   • Took over a critical multi-million AAA multiplatform project at a challenging stage and created a new product vision within two weeks.
   • Led product direction, team alignment and demo delivery for a PC/mobile project; delivered a playable demo in five months with strong external benchmark feedback.
   • Recruited key team members and built a collaborative, high-ownership team around a new vision and execution cadence.
   • Responded to shifting stakeholder requirements by reshaping the product direction toward mobile platforms and delivering a new prototype under tighter constraints.

4. Kool2Play S.A.

   Creative Director / Lead Game Designer

   Warsaw, Poland · April 2020 – September 2023

   • Led the shipped multiplatform title Uragun while rebuilding the studio structure and creative strategy from scratch, hiring senior talent and shaping a 25+ person team.
   • Owned product/creative direction across projects, using playtest feedback, industry metrics and senior-lead input to pivot games and prioritise production backlogs.
   • Introduced Agile workflows, production pipelines and collaboration routines across design, development, external partners and consultants.
   • Led hands-on concept and product direction for an unannounced action co-op game for a major UK publisher.

5. Earlier product, design and leadership roles

   Senior Game Designer, Daftcode / Daftmobile · February 2019 – April 2020
   Designed product vision, gameplay, retention and monetisation systems; led mid-core and hyper-casual prototypes; used UA campaigns, KPIs and A/B testing to assess product direction; contributed to technology/product prototypes in the Daftcode ecosystem.

   Creative Producer, Cherrypick Games · September 2016 – February 2019
   Partnered with studio leadership on product vision and production pipelines; designed systems for My Hospital and My Spa Resort across gameplay, economy, balance, UX and UI flows; used KPIs, A/B testing and qualitative feedback to improve product quality.

   Co-founder / Designer / Producer, We Are Vigilantes · December 2012 – June 2015
   Co-founded an indie studio and shipped Imps in Tokyo, a mobile game praised by TouchArcade; owned concept, design documentation, backlog, production coordination, user communication, PR and post-release iteration.

   Deputy Head / Head of Legal Department, Poland, Orco Property Group · February 2007 – November 2012
   Managed legal operations, strategic advisory, negotiations and crisis communication; built a foundation in risk assessment, structured reasoning, stakeholder alignment and high-stakes decision support.

• MIT Professional Education — AI and ML: Leading Business Growth · October 2025 – March 2026
  Focus: AI/ML initiative lifecycle, business-value framing, opportunity selection, discovery discipline and responsible implementation.
• SWPS University — Postgraduate Studies, Alternative Dispute Resolution · 2007 – 2008
• Ghent University — LL.M., International Law and Legal Studies · 2005
• Lazarski University — Master’s Degree, Law · 2000 – 2005
• Google Cybersecurity Professional Certificate · IDEO U: Leading for Creativity