Software developer on the AI/ML team at SOTI. I lead natural-language app generation in SOTI Snap: plain-English descriptions become working iOS and Android apps through LLM pipelines (RAG, agents, evals), cutting app build time 80%.
Before that I was the sole engineer at a lumber distributor for ~2.7 years and built their whole platform end to end: inventory, quoting (6 hours to 5 minutes), a customer order portal (14% of counter orders moved online), CI/CD (deploys 15 min to 2).
Side projects: llama-wgsl (LLM inference in the browser with hand-written WebGPU kernels, token-exact vs Hugging Face), FlashServe (Llama inference engine with KV cache and speculative decoding), AgentForge (planner-executor-verifier agent framework).
Looking for: AI/applied ML engineer, full-stack or product engineer roles at small, high-ownership teams.
Remote: Yes
Willing to relocate: Yes (US via TN, e.g. SF/NYC)
Technologies: Python, TypeScript, C#, SQL, PyTorch, LLMs (RAG, agents, evals), React, Node, .NET, Postgres, Docker, AWS/GCP, WebGPU
Résumé/CV: https://dilpreetbansi.github.io
GitHub: https://github.com/DilpreetBansi
Email: [email protected]
Software developer on the AI/ML team at SOTI. I lead natural-language app generation in SOTI Snap: plain-English descriptions become working iOS and Android apps through LLM pipelines (RAG, agents, evals), cutting app build time 80%.
Before that I was the sole engineer at a lumber distributor for ~2.7 years and built their whole platform end to end: inventory, quoting (6 hours to 5 minutes), a customer order portal (14% of counter orders moved online), CI/CD (deploys 15 min to 2).
Side projects: llama-wgsl (LLM inference in the browser with hand-written WebGPU kernels, token-exact vs Hugging Face), FlashServe (Llama inference engine with KV cache and speculative decoding), AgentForge (planner-executor-verifier agent framework).
Looking for: AI/applied ML engineer, full-stack or product engineer roles at small, high-ownership teams.