Sergueï Spetschinsky, Ph.D

AI coaching for academics and researchers

Integrating AI into complex intellectual and organisational work — for researchers and scholars in the humanities and social sciences.

Book a free session

I offer AI-powered services for researchers and academics in the humanities and social sciences: individual coaching, workshops and seminars for groups, and consulting for institutions.

At the core of my work is sustained, personalized one-on-one coaching on your actual research, writing, and professional tasks. Each coaching engagement draws on a growing library of AI workflows for academic work. These are tools and methods I research, develop, test, and refine through practice.

The same workflows are available in group workshops, adapted for shared learning. For institutions, I offer consulting on systemic questions at the organizational level.

This offering is constantly evolving: as the technology changes, as I work with clients and learn from them, as new problems surface and new solutions become possible. Many things that are not listed here are probably already feasible or will soon be. When you come with a concrete problem, we often figure out new things together.

Ph.D in Philosophy

10+ years in academia

English · French · German

Based in Berlin

The format and rhythm of our sessions are equally adaptable, meeting you where you are at.

We work directly on tasks and problems you are already facing. No homework, no study sessions. AI tools and workflows are learned by using them on your actual research, and the gains in speed and quality are visible from the first session. Every session produces something you can use immediately.

Depending on what fits your situation, I can teach you a specific workflow and let you run it independently, co-design a system with you over several sessions, build something end-to-end that you receive ready to use, or do something else entirely.

Very quickly if not immediately, you will learn to modify, improve, and eventually create your own AI workflows. Usually, people notice important, and sometimes dramatic, improvements in their work after the very first session.

My goal is to become redundant: that you stop needing me as your coach, and become your very own independent AI system builder.

This is the core of what I do. Everything in this catalog starts from the practice of thinking and writing with AI, whether through individual coaching or group workshops. All the specific workflows listed below involve some form of this partnership. Learning to work with AI as a genuine thinking and writing companion is more of an art than a science, developed through practice and iteration. It is the foundation of using AI as a real research assistant.

1. AI as Thinking and Writing Partner.

A persistent, attentive intellectual companion for your research: sustained, critical engagement available on demand, for as long as you need it.

Argument development, critique & feedback — testing your reasoning, surfacing gaps, generating counterarguments, checking logical consistency, providing structured feedback on any text
Quick research & contextual input — rapidly gathering factual data, cross-disciplinary perspectives, and real-world evidence to ground and enrich the thinking process
Clear expression — translating complex ideas into precise, accessible language for any audience, from specialist peers to general readers
Brainstorming & structured ideation — exploring possibilities, mapping concepts, reframing problems from new angles
Pattern recognition — identifying themes, connections, and structures across your work that are hard to see from inside
Research & writing strategy — planning priorities, structuring the path from notes to finished text, maintaining momentum through the writing process
Conceptual clarification — making implicit assumptions explicit, sharpening distinctions, testing definitions

2. Voice Matching & Style Control.

AI learns how you write — your vocabulary, your sentence rhythms, your disciplinary register — and produces text that sounds like you, not like a machine. The result is a reusable voice profile: a permanent tool you keep and use across any writing task. Multiple profiles can be created for different registers and audiences. The way you write a journal article is not the way you write a grant summary or a blog post. This is a fundamentally different experience from the generic, standardized output that most people associate with AI-generated text.

3. Research Presentations & Slides.

Translating complex research into compelling presentations. This is a skill that directly affects career trajectory and that few academics are trained for.

Simplification — distilling specialist knowledge into clear, accessible language without losing substance
Narrative structure — organizing your argument for oral delivery, which follows different logic than a written paper
Slide design — building clear, effective slides and supporting visuals from your content
Audience adaptation — adjusting depth, terminology, and emphasis for different audiences (specialist conference, interdisciplinary panel, public lecture, funding committee)
Q&A preparation — anticipating questions and preparing responses

4. Public Communication & Outreach.

Making your research accessible beyond the academy: blog posts, op-eds, newspaper articles, Substack and newsletter content, podcast preparation, public lectures, media appearances. AI helps find the right register for non-specialist audiences while preserving the substance of your work.

5. Peer Review.

AI assists on both sides of the review process.

Writing reviews — structuring your evaluation against journal criteria, generating section-by-section assessment, drafting constructive feedback
Responding to reviewers — parsing R&R comments by type, drafting point-by-point responses, planning the revision systematically
Pre-submission review — structured critique that identifies weaknesses before real reviewers see them

6. Research-Grade Translation.

Translations that preserve domain terminology, scholarly nuance, and disciplinary conventions — far beyond what AI consumer translation tools deliver.

Terminology control — enforcing field-specific glossaries across your translations
Quality assurance — structured verification workflows for accuracy and register
Context enhancement — enriching translations with domain-specific knowledge
Post-editing — revision workflows calibrated by text type and purpose
Model selection — systematic evaluation of translation engines for your specific field and language pair

7. Collaborative Writing.

When multiple authors work on a single manuscript, AI helps manage what is usually a painful coordination problem: synthesizing sections written in different styles, harmonizing voice and terminology across contributions, checking that the argument remains consistent from introduction to conclusion, flagging contradictions between sections, and tracking how the manuscript evolves across revision rounds. The result is a more coherent paper, produced with less friction between co-authors.

8. Data Visualization & Figures.

Figures, graphs, and diagrams for papers, presentations, and grant proposals. AI generates, iterates, and refines visualizations from your data, finding the clearest way to present your results for the specific context where they will appear.

9. Thesis & Dissertation Feedback.

Structured, actionable feedback on doctoral chapters and dissertations: detailed, specific, and available on demand.

Argument quality — logical consistency, gap identification, counterargument testing
Language and register — voice, disciplinary conventions, clarity
Citation coverage — source integration and completeness
Defense preparation — anticipating examiner questions and preparing responses

10. Living Personal Archive.

Turning your existing collection of files, notes, and documents — scattered across folders, drives, and forgotten corners — into a living, searchable system. AI makes your personal archive retrievable: old research becomes findable, connections between things you wrote years apart surface, and nothing you have produced is lost anymore.

11. Personal AI-Powered Research Assistant.

Building a centralized, local, portable, and highly personalized research system: a second brain that integrates all the workflows listed above into one environment. It captures, connects, and retrieves everything you produce and encounter in your academic life: reading notes, seminar ideas, supervision notes, research thoughts in transit. It knows your field, your sources, and your arguments. A full prosthetic second brain that evolves with you for the rest of your career — and beyond, providing support across both professional and private life. My own system is the working model. The coaching teaches you to build yours.

12. Literature Review.

Finding, screening, extracting, comparing, and synthesizing sources — without missing critical papers or spending months on the process.

Finding sources — semantic search, citation network exploration, bibliometric field mapping
Screening — AI-accelerated relevance assessment across hundreds of abstracts
Citation analysis — identifying who agrees, who disagrees, and where the field clusters
Gap identification — discovering what is missing in the literature, including unpublished and non-traditional sources
Reading and extraction — batch summarization, framework comparison, structured data extraction
Full pipeline — end-to-end systematic review with transparent, auditable reporting
Living reviews — keeping your literature review current as new publications appear

13. Hypothesis Generation.

AI supports hypothesis development through multiple pathways, matched to your research stage and data availability.

Formulating the question — interactive research question co-creation and refinement
Literature-driven discovery — generating hypotheses from patterns in existing publications
Data-driven discovery — extracting hypotheses from datasets through machine learning and counterfactual analysis
Theory-driven discovery — simulated social experiments, concept-pair prompting
Meta-scientific analysis — mapping trends and gaps across entire fields of research

14. Survey Design & Validation.

AI simulates cognitive interviews and pre-tests your survey with synthetic respondents before you go to the field. Confusing questions, response-scale problems, and acquiescence bias surface in hours instead of weeks of piloting.

15. Interview Protocol Design.

AI identifies bias in your questions, tests alternative wordings, and refines your interview protocol before you sit down with participants, without needing human pilot rounds.

16. Interview Transcription & Processing.

Research-grade transcription of interviews and focus groups, with verification workflows that catch errors and fabricated content that automated tools silently introduce.

Accurate transcription — speaker identification, timestamp alignment, handling of overlapping speech
Hallucination verification — systematic detection of content that the transcription tool invented rather than heard
Formatting for analysis — preparing transcripts for qualitative analysis software, with verbatim and intelligent-verbatim options

17. Data Anonymization.

Removing identifying information from transcripts, field notes, and documents before analysis — preserving analytical value while protecting participants.

Identifying personal information — automated detection of names, locations, dates, and other identifiers in your data
Privacy-safe processing — local pipelines that keep sensitive data off cloud services
Verification testing — systematic checks that anonymization holds and cannot be reversed, including resistance to AI-powered re-identification attempts

18. Historical Document Transcription.

Getting readable, searchable text from handwritten manuscripts, degraded print, and garbled output from optical character recognition software.

Handwritten manuscripts — AI reads directly from images of historical handwriting
Printed historical texts — correcting errors in existing digitized text

19. Document Sorting & Classification.

Thousands of unsorted files with no metadata. A common archival reality.

Semantic classification — AI reads content and categorizes documents by type and topic
Metadata generation — creating structured, searchable metadata for unorganized collections

20. Qualitative Data Analysis.

AI assists across methodological traditions, accelerating the mechanical parts of coding (labeling and categorizing segments of your data) while you stay in control of interpretation.

Tradition-specific workflows — thematic analysis, grounded theory, narrative methods, ethnographic coding, each with approaches adapted to the tradition's logic
Software integration — working with tools like MAXQDA and NVivo, with added consistency checking across your coding
Validation — AI as a second coder for intercoder reliability testing

21. Discourse Analysis.

Applying discourse analysis methods at scale: identifying rhetorical patterns, tracking discursive shifts, and processing large corpora that would be impractical to handle manually.

Critical discourse analysis — systematic analysis of power, ideology, and social structures in texts
Conversation analysis — structural analysis of spoken interaction
Narrative and policy analysis — identifying narrative frameworks in policy documents and public discourse
Stance and frame detection — mapping positions and framing strategies across large collections of texts

22. Published Text Analysis.

Extracting structured, queryable information from large bodies of text: historical archives, literary collections, legal records, contemporary documents.

Information extraction — networks, relationships, timelines, entity mapping
Stylistic analysis — comparing writing styles, authorship questions, register variation
Intertextual analysis — tracing connections, allusions, and borrowings across texts
Topic modeling — discovering thematic structure in large corpora
Rhetorical analysis — mapping argumentative strategies and persuasive structures
Sentiment and stance — tracking attitudes and positions across time or across sources
Close reading support — AI-augmented deep analysis of individual texts, with full source access

23. Small Dataset Augmentation.

When your verified dataset is too small for statistical power, AI-generated predictions — corrected against your ground truth — can bridge the gap without treating AI output as real data.

Synthetic data generation — creating distribution-matched synthetic text or survey data for testing and validation
Few-shot classification — categorizing text reliably from a handful of labeled examples

24. Grant Writing.

AI-assisted support through the entire grant application lifecycle.

The level of my involvement varies: from targeted advice on a specific section you handle yourself, to ready-made workflows and templates you learn to use and develop further, to intensive co-writing where I work alongside you on the content and research strategy itself. The same approach applies to individual scholarships and fellowship applications: personal statements, research plans, CV positioning.

Strategic positioning — analyzing funding trends, assessing your profile against funder criteria
Pre-writing analysis — extracting and mapping funder requirements, mining patterns from successful grants
Writing and revision — AI-assisted drafting with iterative improvement and consistency checking
Pre-submission testing — simulated review panels that stress-test your proposal
Post-rejection — structured diagnosis of what went wrong and how to address it
Funder-specific workflows — tailored processes for specific programs (ERC, DFG, BMBF, Horizon Europe, NSF, NEH, ACLS, Humboldt Foundation, and others)
Compliance — data management plans, regulatory language, reporting requirements
Budget and work plan — timeline construction, budget justification, broader impacts statements

25. Assessment & Feedback for Students.

Giving detailed, specific feedback to every student is one of the most time-intensive parts of teaching, and one where consistency is hardest to maintain across large numbers of submissions. AI applies your standards evenly, generates structured feedback more efficiently, and frees your time for the responses that require your personal judgment.

Grading — consistent application of your criteria across all submissions
Formative feedback — ongoing commentary during the learning process, not just at the end
Writing feedback — targeted, constructive commentary on student writing
Structured peer review — helping students give each other more useful feedback by providing review frameworks and guiding their evaluations
Progress tracking — monitoring learning development over time across assignments

26. Assessment Design for the AI Era.

How do you know your students actually learned something, when AI can produce most of the essays, papers, and assignments they would traditionally submit? Two distinct design challenges, each with its own solutions:

Controlled assessment — designing exams and in-class tasks where AI access is restricted, and where what is tested is genuine understanding rather than memorization
Open assessment — designing assignments where AI use is expected, and where the learning happens through the process of working with AI: critical evaluation, metacognition, judgment about outputs
Hybrid assessment architecture — structuring a complete course assessment plan with different AI access rules for different components, so controlled and open assessments complement each other
Oral assessment design — structuring face-to-face assessments that verify understanding in a format AI cannot shortcut
Process assessment — evaluating how students interact with AI, not just their final product: prompt quality, critical evaluation, revision decisions

27. Course Design & Materials.

Syllabi, reading lists, exercises, slides, presentation materials: the operational infrastructure of teaching. AI accelerates the production of course materials and helps ensure that what you teach, what students do, and what you assess are actually aligned.

Structure and alignment — mapping learning outcomes to activities and assessments systematically
Content creation — reading lists, exercises, case studies, presentation slides, supporting materials
Assessment and activity design — assignments calibrated to work meaningfully with or without AI
Course evaluation analysis — turning hundreds of student feedback comments into actionable redesign priorities
Accessibility review — checking course materials against internationally recognized inclusive teaching standards and generating alternatives for diverse learners

28. Academic Job Market Applications.

Cover letters, CV optimization, teaching and research statements, writing sample selection, each tailored to the specific job ad, institution, and position.

29. Reference Letters.

From your notes to a polished, personalized recommendation draft in minutes.

30. Annual Reports.

CV, publications, and project records compiled into formatted annual reports.

31. Academic Websites.

A working academic website. You walk away with the site, the knowledge to maintain it, and no dependency on a web developer.

32. Email Management.

Triage, prioritization, and draft responses for high-volume academic email. Routine correspondence handled in minutes; you focus on the messages that require your judgment.

33. Ethics & Consent Forms.

Getting informed consent forms and ethics applications through review on the first try, with proper readability, completeness, and regulatory compliance.

Workshops translate the same methodology and workflows used in individual coaching into a group learning format. Participants work hands-on with AI on their own material, learn from each other's questions, and leave with something they can use immediately.

I offer four types of workshops:

AI Literacy for Research.

What AI can and cannot do for academic work: hands-on exploration with real tools, prompt design fundamentals, discipline-specific demonstrations, and an honest discussion of capabilities, limitations, and ethical implications. No prior AI experience required.

Specific AI Workflow Crash Course.

AI fundamentals and a specific advanced workflow taught together in one session. Best suited for groups with shared methodological needs: students from the same field, researchers using similar approaches, or anyone tackling a common challenge. This also works for workflows shared by most academics, such as grant writing (24), literature review (12), peer review (5), career and professional tasks (28–33), or the foundational practice of thinking and writing with AI (1–2). Participants arrive with any level of experience and leave with a working workflow built on their own material.

Advanced AI Research Workflows.

Focused sessions on specific research workflows: literature review, academic writing, peer review, structured thinking, qualitative coding. Participants deepen their existing practice and develop workflows adapted to their own research.

Building Your Personal AI Research Assistant.

The foundations of a personalized AI research system: architecture, capture workflows, retrieval patterns. A powerful tool tailored to your needs that may serve you for the rest of your career, within or beyond academia. Each participant begins building their own system.

All workshops can be held in person or online, and adapted to any format: a focused two-hour session, a full-day intensive, a weekend block seminar, or a semester-long weekly course. All formats can be adapted to any academic audience — from bachelor's students writing their first thesis to senior faculty exploring AI for the first time.

Possible venues and contexts:

—Department study days and retreats
—Graduate school programs and colloquia
—Methods courses and research seminars
—Research group sessions
—Faculty development days
—Conference pre-conference workshops
—Summer and winter schools
—Doctoral school programs
—Student association events
—Library and research support training programs

Institutional consulting addresses systemic challenges that individual coaching cannot reach. These are problems that require coordinated change across a department, faculty, or university.

Assessment & Learning Design.

Helping departments redesign their assessments and course structures for the AI era.

—Collaborative design process with the faculty who actually teach, not a policy document handed down from above
—Evidence-based: which assessment formats hold up when students have AI access (oral exams, process portfolios, AI-integrated assignments)
—Adapted to your discipline and institutional culture
—Concrete implementation plans, not abstract recommendations

Student AI Behavior Analysis.

A structured investigation of how students in your institution actually use AI.

—Empirical findings specific to your context, beyond anecdote and policy anxiety
—Maps what students do, what they believe they do, and what the gap means for teaching
—Concrete recommendations for how to respond

Grant Infrastructure.

Modernizing how your institution supports researchers in writing and managing grants. AI is integrated into the existing grant support infrastructure so it actively strengthens proposals.

—Custom workflows, templates, and extraction pipelines integrated into your existing grant administration
—Coaching and upskilling your grant support staff so they can use and maintain the new tools independently

The following consulting services are under active development:

AI Readiness Assessment. A diagnostic of where your institution stands on AI adoption: faculty comfort levels, student use patterns, existing infrastructure, policy gaps. Produces a concrete roadmap for what to address and in what order.

AI Policy Development. Helping institutions develop clear, workable AI use policies for students, researchers, and teaching staff. What is allowed, what requires disclosure, how to cite AI contributions, how to handle edge cases. Grounded in a philosophical framework of intellectual autonomy and critical judgment, not just compliance.

Faculty AI Development Program. A structured, multi-session program to bring faculty across a department up to speed on AI for their research and teaching. Not a one-off workshop but a sustained development arc. Competence is built progressively, with practice between sessions.

Research Methodology Audit. Reviewing a department's research methodology training for AI-related gaps: where could AI strengthen methods teaching, and which methods courses need updating?

PhD Program Integration. Designing how AI fits into a doctoral program: which competences to develop, which tools to teach, at which milestones, and how to assess AI-related skills.

Publication Strategy. Helping a research group develop an AI-assisted publication pipeline, from draft to submission, including voice consistency, quality control, and journal targeting across the group.

Collaborative research at the intersection of AI, education, and the humanities. Details coming soon.

Autonomy

My goal is your autonomy in all AI matters. I want us to work towards making me redundant as your AI coach.

Understanding

I approach every collaboration by first trying to understand your specific objectives, constraints, and your way of thinking.

Honesty

I will try my best to be of help. But I will also tell you honestly if I think I am not the right person for the job.

Free introductory session

I offer a free introductory session of about 45 minutes. Write to me and we will find a time. We will talk about what you are working on, whether I can help, and you will get a sample of my coaching style.

contact@spetschinsky.com

Sergueï Spetschinsky, Ph.D

About

Services for Individual Researchers

Writing & Thinking

Literature Review

Research Design

Data Processing

Analysis

Grants & Funding

Teaching

Professional & Administrative

Workshops and Seminars