Aneesh Panoli
I work where artificial intelligence meets biology.
Scientist and engineer — deep-learning models of proteins, published genetics research, and a mentor to the next generation of engineers.
Selected work
About
My research asked how a single cell becomes an entire organism. My path since hasn't run in a straight line — molecular genetics, financial markets, machine learning — but it keeps circling back to that question, and the questions next to it that AI is finally making tractable: can we model a mutation's effect in silico, learn the signatures of a cell type, steer a stem cell's fate, build a digital model of human disease?
Here's the belief underneath all of it: the real power of AI isn't handing you the most probable answer — it's that it can write complex code that produces deterministic, reproducible results. Give it the rules of a system and it makes them executable; follow that far enough and the ambition is to codify an entire organism in code. That's why I work where AI meets biology — the rigor of the bench, the discipline of the markets, and the craft of shipping software were all preparation. Some recent work in that direction: reanalyses of public brain data on multiple sclerosis.
Training
- UC DavisPost-Doctoral Fellow · Plant Developmental Biology
- CCMB, HyderabadCSIR Senior Research Fellow
- JNU, New DelhiPh.D. · Molecular Genetics
- Insight Data ScienceHealth Data Science Fellow
Robotics & mentorship
Robotics is where my love of engineering meets teaching. For years I've coached, mentored, and judged competitive robotics — because the best way to grow the field is to grow the people in it.
How a robot drives itself for the first fifteen seconds of a match — odometry, path planning with Choreo, and the routines that string it together, built up from zero.
A beginner's tutorial on FRC shooter design — where to point the turret and how hard to shoot while the robot is driving, built from zero with interactive diagrams.
How the whole FRC control system connects — driver station to roboRIO to motor, and the telemetry that comes back — traced end to end with interactive diagrams.
What the new FIRST robot controller changes for FRC and FTC teams in 2027 — the hardware, the new field coordinates, the driver station, and the tools being retired.
WPILib's new command framework explained for beginners — what a coroutine is, why command bodies read top to bottom now, and what changes coming from Commands V2.
Mentoring one of the most successful FIRST Robotics Competition teams in the world, in Davis, CA.
Coaching teams through design, iteration and competition — headed to VEX Worlds this year.
Coached young teams in robot design and core values — winning at the regional championships.
Serving as a judge on the sport's biggest stage, evaluating engineering design and problem-solving.
On the Scientific Review Committee, safeguarding the integrity of student research projects.
Simulation
Biology is hard to experiment on and easy to fool yourself about. Simulation is how I check the reasoning before trusting it — building a model of what the data should look like when nothing interesting is happening, so a real signal has something honest to stand against.
Twenty lines of Python grow a cowpea for thirty-five simulated days — then measure its height, leaf count and leaf area, and see why those numbers are free in software and cost a destructive harvest in a field.
Reads
Science, markets, machine learning, and the mind — a few of the books that shaped how I think.
Neural art
Experiments in neural style transfer.
