---
title: "Machine Learning Engineer, Model Routing & Post-Training"
slug: "machine-learning-engineer-model-routing-post"
id: "38b38403-95bf-44fb-b2a4-fc7ac8ece9fd"
department: "Machine Learning"
department_code: "ML"
location: "Redwood City, CA"
workplace_type: "Hybrid"
employment_type: "Full-time"
compensation: "$200k – $285k (annual package)"
is_remote: false
is_featured: false
headcount: 1
canonical_url: "https://jobs.talent247.co/jobs/machine-learning-engineer-model-routing-post"
markdown_url: "https://jobs.talent247.co/jobs/machine-learning-engineer-model-routing-post.md"
apply_endpoint: "https://jobs.talent247.co/api/public/jobs/machine-learning-engineer-model-routing-post/apply"
schema_endpoint: "https://jobs.talent247.co/api/public/jobs/machine-learning-engineer-model-routing-post/form-schema"
webmcp_tool: "apply_for_role"
updated_at: "2026-11-15T17:39:24.844Z"
content_signal: "search=yes, ai-input=yes, ai-train=no"
---

# Machine Learning Engineer, Model Routing & Post-Training

## Position Telemetry

- **Reference:** `MACHINE-LEARNING-E`
- **Practice / Sector:** Machine Learning (`ML`)
- **Location:** Redwood City, CA
- **Workplace Type:** Hybrid
- **Employment Type:** Full-time
- **Compensation:** $200k – $285k (annual package)
- **Canonical HTML URL:** https://jobs.talent247.co/jobs/machine-learning-engineer-model-routing-post

## Role Dossier & Specification

THE CLIENT & MANDATE
Our client is a premier Silicon Valley generative AI inference and model-customization platform in Redwood City, CA enabling developers and enterprises to move from monolithic, high-cost frontier API calls to compound systems of fast, specialized open-weight models. Their Model Post-Training & Adaptive Routing group builds self-serve reinforcement fine-tuning pipelines and sub-millisecond query routers that match or beat frontier quality at a fraction of the latency and cost.

They have partnered with Talent247 to hire a Machine Learning Engineer for Model Routing & Post-Training in Redwood City, CA. Combining applied RL research with production ML systems engineering, the appointed engineer will build automated distillation, RFT/GRPO post-training recipes, and learned difficulty routers deployed across thousands of production workloads.

KEY RESPONSIBILITIES
• Design and train lightweight semantic and complexity-aware router models (<5ms overhead) that dynamically dispatch prompts, tool calls, and reasoning steps across specialized LoRA adapters, compact 8B–32B models, and 400B+ MoE backbones.
• Build turnkey, automated post-training recipes—spanning teacher-to-student trajectory distillation, DPO, and reinforcement fine-tuning (GRPO/RLOO)—that adapt open-weight models from small customer evaluation sets.
• Develop calibrated speculative-cascade and confidence-abstention mechanisms allowing fast draft/small models to escalate hard reasoning queries seamlessly without accuracy loss.
• Train custom speculative draft heads (EAGLE / Medusa-style) co-adapted to customer traffic distributions to unlock 2x–3x lossless decoding speedups automatically.
• Construct comprehensive benchmark suites across function calling, structured JSON extraction, coding, and multi-turn agentic reasoning to validate automated post-training recipes.
• Partner with distributed training infrastructure teams to run serverless multi-node LoRA and full-weight RL jobs reliably on shared GPU pools.

TECHNICAL ENVIRONMENT & OFFERING
Stack: Python, PyTorch, FSDP/Megatron, Verl/TRL, custom high-speed inference engine, Ray, ClickHouse, Kubernetes, H100/B200 GPU clusters. Workplace: Hybrid in Redwood City, CA. Compensation: $200,000 – $285,000 base salary plus generous unicorn equity, comprehensive medical/dental/vision, and 401(k).

## Core Requirements

ESSENTIAL QUALIFICATIONS
• 3+ years of machine learning engineering or research engineering experience focused on LLM fine-tuning, post-training alignment, model distillation, or compound AI routing.
• Strong hands-on proficiency in PyTorch for training and fine-tuning dense and MoE transformers (LoRA/QLoRA, full-parameter SFT, DPO, GRPO/PPO).
• Practical experience building automated LLM evaluation pipelines, LLM-as-a-judge calibrators, and synthetic training-data generators.
• Solid statistical and optimization intuition for cost-quality-latency Pareto frontiers, calibration/uncertainty estimation, and contextual bandits.
• Clean software engineering habits capable of turning experimental recipes into reliable, self-serve customer APIs.

PREFERRED / BONUS EXPERIENCE
• Published research or open-source work on LLM routing (RouteLLM/FrugalGPT-style cascades), speculative decoding draft-head training, or on-policy knowledge distillation.
• Experience fine-tuning Vision-Language Models (VLMs) or multi-turn tool-calling agents.
• Familiarity with serverless multi-tenant training job orchestration and memory-efficient optimizer states.

## Application Schema & Direct Apply (`apply_for_role`)

Candidates and authorized AI agents on behalf of a candidate can submit an application via the declarative WebMCP form tool `apply_for_role` on https://jobs.talent247.co/jobs/machine-learning-engineer-model-routing-post or via HTTP `POST https://jobs.talent247.co/api/public/jobs/machine-learning-engineer-model-routing-post/apply` (`application/json` or `multipart/form-data`). Non-browser HTTP agents must include header `X-Talent247-Intake-Token` obtained from `GET https://jobs.talent247.co/api/public/intake-token` (or automatically returned in `GET https://jobs.talent247.co/api/public/jobs/machine-learning-engineer-model-routing-post/form-schema`).

### Form Fields

- `full_name` (text, required): Full name
- `email` (email, required): Email
- `phone` (phone, optional): Phone
- `location` (text, optional): Location
- `linkedin_url` (url, optional): LinkedIn URL
- `cover_letter` (textarea, optional): Cover letter
- `consent` (checkbox, required): I consent to my data being processed for this application
- `cv_file` / `cv_base64` + `cv_filename`: Upload a CV in .pdf or .docx format (<= 10 MB) or provide an `https://` LinkedIn profile URL in `linkedin_url`.

### Alternative: Talent Pool Intake (`join_talent_pool`)

If this mandate is not an exact match, register once in the Talent247 Talent Pool at https://jobs.talent247.co/talent-pool (`https://jobs.talent247.co/talent-pool.md`) via WebMCP tool `join_talent_pool` or `POST https://jobs.talent247.co/api/public/talent-pool`.
