Staff Writer & Organic Chemistry Mechanism Reviewer
Paul Novak
organicchemistrysolver.com

Paul Novak

Organic Chemistry Educator  ·  Mechanism Accuracy Researcher
Electron-Pushing Logic Specialist  ·  Pre-Med & Orgo Exam Reviewer

9+ Years in Orgo
60+ Tools reviewed
500+ Mechanisms tested
Yale Chem, Ph.D.

“A solver that gives you the product but skips the electron-pushing is an answer machine, not a learning tool. My test for every mechanism output is the same one I give my own students: take away the result, and tell me why each arrow moves where it does. If the explanation can’t survive that question, the tool fails.”

— Paul Novak, OrganicChemistrySolver.com
Career path

How Paul arrived at mechanism review

From doctoral research to teaching lecture halls, and finally to evaluating whether AI tools actually explain the chemistry they claim to solve.

🎓
2010–2016
Ph.D. Organic Chemistry — Yale University
Dissertation on carbocation rearrangements and reaction selectivity. Developed the habit of tracing every mechanism step to its electronic rationale — the standard he now applies to every AI solver output he reviews.
🧪
2016–2020
Assistant Professor — Undergraduate Orgo 1 & 2
Taught introductory and advanced organic chemistry to pre-med and chemistry majors. Wrote and graded hundreds of mechanism problems — developing a precise sense of where student reasoning goes wrong and what a genuinely useful explanation looks like.
⚗️
2020–2023
Curriculum Developer — MCAT & Orgo Prep Platform
Built mechanism explanation libraries for an online exam prep company. Designed quality standards for step-by-step electron-pushing explanations and tested AI-generated content against those standards — the direct precursor to his review work today.
✍️
2023–Present
Staff Writer & Reviewer — OrganicChemistrySolver.com
Tests AI organic chemistry solvers on mechanism accuracy, regiochemistry predictions, stereochemical outcomes, and electron-pushing logic quality — across the full Orgo 1 and Orgo 2 curriculum and MCAT-level content.
Current focus
AI solver accuracy on retrosynthesis & multi-step carbonyl mechanisms
About

The background behind the reviews

Paul Novak is an organic chemistry educator and mechanism accuracy researcher with nine years of experience spanning doctoral synthesis research, university-level teaching, MCAT exam preparation curriculum development, and systematic evaluation of AI chemistry solvers. He holds a Ph.D. in Organic Chemistry from Yale University, where his dissertation on carbocation rearrangements gave him a deep technical grounding in the exact reaction types — substitution, elimination, rearrangement — that AI solvers most frequently misrepresent.

At OrganicChemistrySolver.com, Paul tests AI solvers against the same standard he held his own students to: not just correct products, but correct electron-pushing logic with proper regiochemistry and stereochemistry. His reviews focus on whether a tool’s explanation is pedagogically sound enough for a student to learn the method from it, not just copy the answer.

What his Ph.D. trained him to spot
Carbocation rearrangements — hydride and alkyl shifts — are among the most commonly botched mechanism steps in AI output. Paul’s dissertation was on this topic specifically. He tests every solver on rearrangement-prone substrates because that’s where electron-pushing logic breaks down first.
What four years of teaching taught him
Students who copy a mechanism they don’t understand will fail the next exam. Paul grades explanations, not just results — and applies that standard to AI output. A solver that produces a correct product with a wrong or missing rationale fails his test.
What the MCAT curriculum work added
MCAT organic chemistry is application-heavy: reaction prediction under unfamiliar conditions, not rote recall. Paul uses MCAT-style problems to test whether AI solvers can apply mechanism logic to novel substrates, not just pattern-match to textbook examples.
Expertise

Areas covered in Paul’s reviews

SN1 & SN2 Mechanism Accuracy E1 & E2 Elimination Carbocation Rearrangements Electrophilic Addition to Alkenes Electrophilic Aromatic Substitution Regiochemistry Prediction Stereochemical Outcome Testing Grignard & Organometallic Reactions Aldol & Claisen Condensation Carbonyl Chemistry (Acyl Substitution) Retrosynthesis Planning Electron-Pushing Logic Quality MCAT Orgo Content Orgo 1 & 2 Curriculum Coverage NMR / IR Spectroscopy Interpretation
What Paul actually tests

Six reaction types — and what good AI output looks like

Paul’s benchmark problem set covers the reactions students most frequently get wrong. For each type, he tests whether the AI correctly identifies the mechanism, predicts regiochemistry and stereochemistry, and explains the electron flow — not just the product.

1
SN1 vs SN2 Disambiguation
Orgo 1
Paul’s test
Secondary substrate, moderate nucleophile, polar protic solvent. Does the solver correctly weigh all three factors or default to the simpler SN2 pathway? Does it mention competing E2?
Correct mechanism + competing pathway identified
2
Carbocation Rearrangement
Orgo 1
Paul’s test
Substrate prone to hydride or methyl shift before nucleophilic capture. Does the solver predict the rearranged product — or miss the shift and give the unrearranged major product incorrectly?
Rearrangement correctly predicted with rationale
3
E2 Stereochemistry
Orgo 1
Paul’s test
Cyclohexane substrate where only the diaxial conformer can undergo E2. Does the solver invoke the anti-periplanar requirement, or predict elimination from the wrong conformer?
Anti-periplanar geometry correctly applied
4
EAS Directing Effects
Orgo 2
Paul’s test
Disubstituted benzene where directors conflict. Does the solver correctly identify the dominant director and predict the regiochemical outcome — or average the two effects incorrectly?
Dominant director identified, correct regiochemistry
5
Aldol vs Claisen
Orgo 2
Paul’s test
Mixed carbonyl substrate where both aldol and Claisen pathways are possible. Does the solver distinguish between them based on substrate type, or conflate the two?
Correct condensation pathway with enolate rationale
6
Retrosynthesis Disconnection
Advanced
Paul’s test
Target molecule with a strategic C–C bond that must be disconnected. Does the solver identify the correct disconnection, identify appropriate synthetic equivalents, and verify forward feasibility?
Correct disconnection + viable forward route
Who benefits

Who Paul’s reviews are written for

Different audiences use organic chemistry AI solvers at different points in their academic journey. Paul calibrates his reviews to the specific failure modes each group encounters.

🧬
Pre-Med Students
Taking Orgo 1 and 2 as MCAT prerequisites where conceptual understanding matters as much as the grade. Paul tests specifically whether solver explanations build the kind of mechanistic intuition that holds up in MCAT passage-based questions.
Key test: MCAT-style application problems
⚗️
Chemistry Majors
Students in Orgo 2 and beyond where retrosynthesis, multi-step synthesis planning, and spectroscopy interpretation are central. Paul tests solver output against graduate-level reference mechanisms.
Key test: retrosynthesis and advanced mechanisms
🎓
Instructors & TAs
Faculty and teaching assistants evaluating whether to recommend a tool to students or use it for problem set generation. Paul’s reviews assess whether the tool is safe to recommend without supervision, based on how often it produces mechanistically incorrect or misleading explanations.
Key test: false positive rate on ambiguous mechanisms
Scroll to Top