Nima Manaf

Gazetteer · 38.70° N → 52.09° N

Nima ManafPhD

Building a Future to Believe In

My work sits between machine learning, optimization, finance, and operations research. I care about models that survive contact with constraints, teams, and real decisions.

Read the articles See timeline

Latest essay De DAW of the Future

Current work Data Analytics Lead
Current base Netherlands
Current team Rabobank / Value Chain Finance Portfolio

AI Scientist

Portrait of Nima Manaf
Nima Manaf, PhD Netherlands 52.09° N, 5.12° E
Plate 01 2022 - Present Netherlands 52.09° N, 5.12° E

Rabobank

The current chapter is Rabobank, value chain finance, and research habits translated into production decisions.

  • The work now sits closer to decisions: portfolio monitoring, credit risk, model governance, and analytics that teams actually use.
  • The academic habits are still there, but the output has to survive operational, regulatory, and business constraints.

Bearings carried forward

Research moved into finance, risk, and portfolio analytics.

Current chapter

Data Analytics Lead

Rabobank / Value Chain Finance Portfolio

Rabobank office buildings on Croeselaan in Utrecht
Rabobank Headquarters / Croeselaan 18, 3521 CB Utrecht, Netherlands
Plate 02 2020 - 2022 Netherlands 51.44° N, 5.48° E

Postdoc at TU/e

The postdoc years were a proud way to give back to academia: helping bachelor's and master's students graduate, building lasting friendships and a wider research network, and joining Ivo Adan's group, a funny coincidence because his name appeared first in the references of my PhD thesis.

  • Helping bachelor's and master's students graduate became one of the parts I am proudest of.
  • Eindhoven also built a wider academic network and friendships that stayed beyond the projects.
  • The coincidence still makes me smile: I joined Ivo Adan's group after seeing his name first in my PhD thesis references.
Atlas Building at Eindhoven University of Technology
Fig. 02 · TU/e Atlas Building Image: Honeywell

Bearings carried forward

A proud giving-back chapter: student graduations, friendships, and the Ivo Adan coincidence.

Plate 03 2020 Turkey 41.01° N, 28.98° E

Final PhD stretch

I returned to Istanbul to write and defend the dissertation: Modelling and Control of Production Systems Based on Observed Inter-event Times: An Analytical and Empirical Investigation.

  • The return to Istanbul was about closing the loop: writing, defending, and turning years of work into a finished thesis.
  • The dissertation studied how observed inter-event times can be used to model and control production systems.
  • PhD Academic Excellence Award recipient.
Dissertation on arXiv
Koc University Rumelifeneri campus overlooking the Black Sea
Fig. 03 · Koc University, Rumelifeneri Campus Image: apply.ku.edu.tr

Bearings carried forward

Dissertation defended, PhD Academic Excellence Award, and the research bridge into applied decision systems.

Plate 04 2019 - 2020 Germany 48.78° N, 9.18° E

Bosch Center for Artificial Intelligence

A year at Bosch AI working on scheduling, reinforcement learning, search, and industrial data.

  • Bosch put the research inside industrial constraints: scheduling, search, reinforcement learning, and semiconductor data.
  • It made the work more concrete, because the algorithms had to meet factory-scale messiness.
Bosch research campus building in Renningen
Fig. 04 · Bosch Renningen Campus Image: Wikimedia Commons

Bearings carried forward

One year inside industrial AI.

Plate 05 2014 - 2019 Turkey 41.01° N, 28.98° E

Koc University, MSc to PhD

Istanbul became the long academic chapter: a full-scholarship MSc and then a full-scholarship PhD at Koc University.

  • MSc in Operations and Information Systems, 2014 - 2016, on full scholarship.
  • The master's work turned into a full-scholarship doctoral route through operations research, production systems, and data-driven modeling.
  • This is where I started treating uncertainty as something to model carefully, not smooth away.
Koc University Rumelifeneri campus overlooking the Black Sea
Fig. 05 · Koc University, Rumelifeneri Campus Image: apply.ku.edu.tr

Bearings carried forward

Full-scholarship MSc and PhD years that turned operations research into a doctoral path.

Plate 06 2010 - 2014 Iran 35.69° N, 51.39° E

BSc at Sharif University of Technology

Tehran meant a full-scholarship BSc in Industrial Engineering at Sharif University of Technology, often described as Iran's MIT.

  • BSc in Industrial Engineering, 2010 - 2014, on full scholarship.
  • Exceptional Talent Student distinction: top 0.1% in the National University Entrance Exam.
  • Industrial engineering gave me the language of systems, constraints, queues, and trade-offs.

Bearings carried forward

Full-scholarship BSc at Sharif, exceptional talent distinction, and the first formal step into systems thinking.

Plate 07 Beginning East Azerbaijan, Iran 38.70° N, 46.55° E

Born among mountains

I was born in Qaradag, a region of Iranian Azerbaijan held on every side by mountains.

  • Look left and there are mountains. Look right, ahead or behind, and there are more, some bigger than the others. They make you wonder whether what is up there is the zirve, the summit.
  • A mountain leaves you two choices: surrender or fight. If you surrender, you know the end is close. So the mountains make two kinds of people: the savaşanlar, the ones who fight, and the ones who give up.
  • That Qaradag energy shows in some of the people of these mountains. Babak held them for twenty years against the invading Abbasid armies; his fortress still stands in Qaradag and is seen as a symbol of freedom.
  • Sattar Khan, the son of a Qaradag family, held Tabriz for Iran's constitution in 1908–09, against the Shah's armies and the Russian power behind them.
Babak Castle on its mountaintop in the Qaradag mountains
Fig. 07 · Babak Castle, Qaradag Image: Wikimedia Commons

Bearings carried forward

Mountains make you either surrender or fight. Qaradag raises savaşanlar.

Current Practice

Current Work

The current chapter sits at the intersection of financial risk, optimization, machine learning, and decision systems used in practice.

Financial Risk and Portfolio Analytics

Portfolio monitoring, challenger modeling, regulation-facing analysis, and decision support for value chain finance and credit risk workflows.

Credit riskValue chain financeMonitoringModel governance

Learning and Optimization

Reinforcement learning, evolution strategies, graph-based modeling, and optimization methods for environments with delayed feedback and hard constraints.

Reinforcement learningEvolution strategiesGNNsStochastic optimization

Operational Systems

Production planning, inventory control, scheduling, and data-driven performance analysis for industrial systems that have to work outside the lab.

ManufacturingSchedulingInventoryOperations research

Selected Work

A few threads that stayed with me.

Plate 01 · Current work

Value Chain Finance Analytics

Leading analytics initiatives at Rabobank with a focus on portfolio performance, risk, and decision quality.

Open

Plate 02 · Research thread

Learning to Control Inventory Management Systems

One gradient-free recipe — CMA-ES over small, interpretable policies — learns competitive policies across ten classical inventory-control problems.

Open

Plate 03 · PhD work

Production Systems from Event Data

Doctoral work on modeling production systems from observed inter-event times and using those models for better control.

Open

Writing

Selected Writing

Research notes, visual experiments, and technical essays.

All articles

Oct 29, 2026

De DAW of the Future

For more than thirty years the digital audio workstation has been a studio you must learn before it will serve you. Now that sound is abundant and agents can edit, the scarce thing is no longer the operation but the intention. The next DAW is an instrument that people and agents shape together: the instrument supplies the senses, the work keeps its memory, and the ear keeps the verdict.

Jul 11, 2026

Fourteen Journeys: How Value Chains Reorganize Around Human Wants

Fourteen industries took fourteen routes, and one order explains them all: human nature changes slowest, the wants it produces outlast every form that serves them, and technology keeps moving what is scarce while each value chain reorganizes around the same want. An atlas of where fourteen chains stand in 2026.

Jun 14, 2026

The Selection Pressure: How Cost Decides Which Value Chains Survive

A value chain is the path a human want travels from unmet to met, and at every link sits a cost. The want is the constant; cost does the selecting; and when one cost falls, the rent relocates to whichever cost has not yet fallen. Run it forward and the costs fall from the outside in, until the last ones standing are human.

Contact

If the problem is real and difficult, I am interested.

I am especially drawn to work that combines data, uncertainty, and operational constraints.

Email nima.manaf8@gmail.com LinkedIn Professional profile GitHub Code and experiments Scholar Publications and citations