The Nudge Group
Founding Software Engineers - Full-Stack/AI & Front-End (Patient Experience)
The Nudge Group, San Francisco, California, United States, 94199
On-site, Bay Area | Early-stage, high-intensity | Equity on offer
Our client is a stealth, pre-seed oncology AI startup building an AI “second-opinion” engine that converts raw EHR data into clear, personalized cancer treatment plans. Single-digit team, massive mission, real ownership.
Work setup (non-negotiable):
On-site in the Bay Area, 5 days/week + 1 day WFH (6-day cadence).
We over-communicate the intensity to avoid mismatches, this is true 0→1 startup building.
Why join
Immediate impact:
You’ll be ~hire #7 shaping architecture, product, and culture.
Serious data moat:
Exclusive, multimodal clinical data (notes, radiology, pathology, genomics) at meaningful scale.
Frontier stack:
Agentic systems, tool-augmented reasoning, RAG/KAG grounding, robust eval and safety scaffolds.
Mission > hype:
Ship software that helps clinicians and patients make life-changing decisions.
Open roles
1) Founding Full-Stack / AI Engineer
What you’ll do (first 90 days)
Orchestrate
planner–executor–critic
agent workflows across retrieval, tools, and safety gates.
Own
post-training
of frontier LLMs (RL, RLHF/RLAIF, LoRA/PEFT); build clinical
reward models .
Expand the
evaluation harness
(synthetic vignettes, adversarial probes, offline metrics).
Prototype
temporality-aware embeddings
for longitudinal EHR reasoning.
Ship fast end-to-end: services, pipelines, deployment, monitoring.
You bring
0→1 builder who’s shipped real systems; high agency by default.
Backend/infra:
Python ,
PyTorch/JAX ,
FastAPI/Django , CI/CD, observability; distributed data/eval experience.
Comfortable jumping into
React/TypeScript/Tailwind
when needed.
Bonus: DeepSpeed/FSDP, RL/RLAIF, healthcare privacy/compliance.
2) Front-End Engineer (Patient Experience)
What you’ll do (first 90 days)
Own
patient & clinician UIs
end-to-end (intake, dashboards, timelines, recommendations, handoffs).
Architect the design system
(React + Tailwind) with accessibility and motion baked in.
Stand up secure
Python (FastAPI)
endpoints for patient data intake and model outputs.
Build feedback loops that improve UX and model performance with every interaction.
You bring
Craft-focused product engineer; accessible, production-grade UI is your default.
React, TypeScript, Tailwind/modern CSS ; comfort with Python APIs; CI/CD, AWS/GCP, observability.
Startup DNA: bias to action, user-obsessed, comfortable with ambiguity.
Bonus: LLM UX, RAG/KAG, agent patterns, HIPAA/SOC-2 familiarity.
What we offer
Competitive salary + equity on offer
Top-tier health benefits, hardware & learning budget, conference stipend
Visa & relocation support (O-1/H-1B)
Lightning-fast, founder-led process (often under a week)
If you thrive in high-ownership, high-tempo environments and want your work to matter—let’s talk.
#J-18808-Ljbffr
Our client is a stealth, pre-seed oncology AI startup building an AI “second-opinion” engine that converts raw EHR data into clear, personalized cancer treatment plans. Single-digit team, massive mission, real ownership.
Work setup (non-negotiable):
On-site in the Bay Area, 5 days/week + 1 day WFH (6-day cadence).
We over-communicate the intensity to avoid mismatches, this is true 0→1 startup building.
Why join
Immediate impact:
You’ll be ~hire #7 shaping architecture, product, and culture.
Serious data moat:
Exclusive, multimodal clinical data (notes, radiology, pathology, genomics) at meaningful scale.
Frontier stack:
Agentic systems, tool-augmented reasoning, RAG/KAG grounding, robust eval and safety scaffolds.
Mission > hype:
Ship software that helps clinicians and patients make life-changing decisions.
Open roles
1) Founding Full-Stack / AI Engineer
What you’ll do (first 90 days)
Orchestrate
planner–executor–critic
agent workflows across retrieval, tools, and safety gates.
Own
post-training
of frontier LLMs (RL, RLHF/RLAIF, LoRA/PEFT); build clinical
reward models .
Expand the
evaluation harness
(synthetic vignettes, adversarial probes, offline metrics).
Prototype
temporality-aware embeddings
for longitudinal EHR reasoning.
Ship fast end-to-end: services, pipelines, deployment, monitoring.
You bring
0→1 builder who’s shipped real systems; high agency by default.
Backend/infra:
Python ,
PyTorch/JAX ,
FastAPI/Django , CI/CD, observability; distributed data/eval experience.
Comfortable jumping into
React/TypeScript/Tailwind
when needed.
Bonus: DeepSpeed/FSDP, RL/RLAIF, healthcare privacy/compliance.
2) Front-End Engineer (Patient Experience)
What you’ll do (first 90 days)
Own
patient & clinician UIs
end-to-end (intake, dashboards, timelines, recommendations, handoffs).
Architect the design system
(React + Tailwind) with accessibility and motion baked in.
Stand up secure
Python (FastAPI)
endpoints for patient data intake and model outputs.
Build feedback loops that improve UX and model performance with every interaction.
You bring
Craft-focused product engineer; accessible, production-grade UI is your default.
React, TypeScript, Tailwind/modern CSS ; comfort with Python APIs; CI/CD, AWS/GCP, observability.
Startup DNA: bias to action, user-obsessed, comfortable with ambiguity.
Bonus: LLM UX, RAG/KAG, agent patterns, HIPAA/SOC-2 familiarity.
What we offer
Competitive salary + equity on offer
Top-tier health benefits, hardware & learning budget, conference stipend
Visa & relocation support (O-1/H-1B)
Lightning-fast, founder-led process (often under a week)
If you thrive in high-ownership, high-tempo environments and want your work to matter—let’s talk.
#J-18808-Ljbffr