Moughees Ahmed

Currently working on:

Doing ML research + engineering at Rime Labs - text-to-speech systems, founding engineer, raised 3M+, <150ms inference, serving large enterprise customers

Past experiences:

Engineer #4 (Raised 10M+) at Loop Financial, company has processed over 1B in transactions.
Founding Engineer at Column: Raised $2M, built expert-driven social network, hired team. Incubated at Social Impact Capital.
Generative AI Research (Speech/Deepfakes) at CSALT - LUMS Pakistan(NLP Research Lab) - research was funded by Facebook, under a CMU PhD. Worked on cutting-edge TTS/ASR models for low-resource languages.

Broadly interested in:

Discovering new knowledge using LLMs: how do we accelerate scientific discovery?
- Can we use search (MCTS) + CoT + dynamic sampling on existing models to accelerate scientific discovery? (Run GPT-5 for months on end to search for a new drug)
- It is likely that a lot of the knowledge needed for a frontier discovery is already published. Instead of next token prediction, can we do next-experiment-prediction?
Maximizing performance of SLMs
Computational genetics + bioML:
- Cost of intelligence and sequencing is dropping, what patterns are we not matching?
- What can we learn about prehistory?
Post AGI* questions:
- Uber and Airbnb maximized the utilization of cars and houses on the back of mobile + internet, what is the AGI equivalent?
- If capital allocation improves 10x, what new asset classes open up?
- Can we create synthetic AGI backed assets as risk-free securities? As liquid and stable as treasury bonds.
*AGI: Artificial General Intelligence: I define it as a drop-in remote worker, mostly digital for now.
Prediction markets: How do we improve liquidity and scale them?
Lowering housing costs in Toronto.
If Starship works, can we airdrop excess food supply from space? (land directly where needed most)

BioML Research: Developed machine learning models for read-level DNA analysis, predicting specific gene presence with over 90% accuracy prior to full gene sequencing. Collaborated with an infectious disease laboratory at Emory University and a UIUC PhD. Unfortunately, our publication plans were preempted by a similar study in Nature.
Organized a builders hackathon with sponsorships from Chroma, Replit and Cohere - gave away 5K in cash prizes. Goal was to mobilize Toronto builders.

Always open to deeper conversations!