AI Engineer III at Gopuff

United States

Gopuff Logo
Not SpecifiedCompensation
Senior (5 to 8 years)Experience Level
Full TimeJob Type
UnknownVisa
Technology, E-commerceIndustries

Requirements

  • Expertise in context engineering, RLHF/RLVR, and low-latency serving for agentic AI systems
  • Experience leading technical architecture, standards, and evaluation strategies connecting research to production
  • Proficiency in multi-modal context integration (temporal, spatial, behavioral) and real-time grounding with dynamic constraints
  • Knowledge of retrieval systems, memory policies, context schemas, and data contracts
  • Familiarity with declarative prompt/program compilation tools (e.g., DSPy)
  • Experience designing multi-agent orchestration patterns (e.g., LangChain/LangGraph, CrewAI, AutoGen, LlamaIndex)
  • Skills in supervised fine-tuning, data curation, synthetic data generation, QA, golden sets, and rubric/pairwise evaluations
  • Expertise in reasoning architectures including planning, tool selection, reflection, and uncertainty-aware decision-making
  • Knowledge of parameter-efficient adaptation strategies (LoRA/QLoRA, text-to-LoRA)
  • Proficiency in RLHF/RLVR pipelines, preference data loops, scalable oversight, and guardrails
  • Experience with policy optimization methods (DPO/PPO/GRPO/GSPO) and advanced techniques (constrained optimization, regularized objectives, KL-control)
  • Understanding of offline-to-online correlation via counterfactual/IPS/DR estimators and stress testing
  • Skills in interpretability, controllability, alignment verification, safeguards against reward hacking, and distributional robustness
  • Knowledge of privacy-preserving methods (data minimization, federated/on-device learning)
  • Expertise in low-latency inference architectures (quantization, caching, batching, streaming) with resilient fallbacks and red-teaming
  • Experience building evaluation frameworks coupling offline metrics with online performance and safety
  • Ability to use APIs for high-fidelity data augmentation to improve grounding and suggestions
  • Capability to design experiments, define success criteria, and iterate with Engineering and Data Science teams
  • Strong skills in translating product goals into technical milestones, documentation, incident response, and learning playbooks
  • Mentoring experience and ability to raise standards in design quality, reproducibility, and ethical rigor

Responsibilities

  • Lead technical efforts across context engineering, RLHF/RLVR, and low-latency serving
  • Define architecture, standards, and evaluation strategy connecting research to real-world impact
  • Mentor colleagues and influence cross-functional roadmaps
  • Ship production systems delivering measurable customer and business outcomes
  • Set strategy for context engineering to maximize precision/recall of order metrics across sessions, households, locales, and time
  • Architect multi-modal context integration and real-time grounding with dynamic constraint satisfaction
  • Establish retrieval freshness, geo/time-aware constraints, memory policies, context schemas, and data contracts
  • Champion declarative prompt/program compilation for systematic LLM behavior
  • Design multi-agent orchestration patterns for robust emergent reasoning
  • Lead supervised reasoning-centered fine-tuning with data curation, synthetic data, QA, golden sets, and evals
  • Own reasoning architecture and evaluation strategy for robust, low-latency outcomes
  • Drive parameter-efficient adaptation strategies with criteria for specialization vs. generalization
  • Architect RLHF/RLVR pipelines, preference data loops, scalable oversight, and guardrails
  • Own policy optimization strategy with safety considerations
  • Ensure robust offline-to-online correlation via estimators and stress tests
  • Establish interpretability, controllability, and alignment verification practices
  • Develop safeguards against reward hacking, unsafe exploration, and ensure content policy compliance
  • Advance privacy-preserving methods with privacy-by-design
  • Architect low-latency, cost-efficient inference with resilient fallbacks and red-teaming
  • Build eval frameworks coupling offline/online metrics with promotion gates
  • Use APIs for data augmentation strengthening grounding and suggestions
  • Partner with Engineering and Data Science on experiments, success criteria, and iteration
  • Translate product goals into technical milestones with documentation and playbooks
  • Mentor colleagues and elevate design quality, reproducibility, and ethical rigor

Skills

DSPy
LangChain
LangGraph
CrewAI
AutoGen
LlamaIndex
RLHF
RLVR
fine-tuning
context engineering
multi-agent orchestration
synthetic data generation

Gopuff

Digital delivery service for convenience items

About Gopuff

GoPuff is a digital delivery service that specializes in providing convenience items, snacks, and beverages directly to consumers. Customers can easily place orders through the GoPuff website or app, and their selected products are delivered straight to their doorsteps. The service offers a wide range of items, from candy bars to alcoholic beverages, catering to those who want quick access to everyday goods. Revenue is generated through product sales and delivery fees, with promotional deals available to attract more customers. What sets GoPuff apart from its competitors is its extensive product selection and commitment to fast delivery, aiming to meet the needs of consumers looking for convenience.

Philadelphia, PennsylvaniaHeadquarters
2013Year Founded
$4,817.2MTotal Funding
CONVERTIBLECompany Stage
Consumer GoodsIndustries
1,001-5,000Employees

Benefits

Comprehensive medical, dental, and vision insurance
Optional FSAs and HSA plans
401(k)
Commuter benefits
Supplemental employee, spouse and child life insurance to all eligible employees
Gopuff employee discount
Career growth opportunities
Internal rewards programs
Annual performance appraisal and bonus

Risks

Increased competition from DoorDash could impact Gopuff's market share in gift card delivery.
'Santa Hours' may lead to higher operational costs without guaranteed returns.
Aldi price match could strain profit margins if not managed carefully.

Differentiation

Gopuff offers 24/7 delivery with 'Santa Hours' in London, enhancing convenience.
Partnership with Blackhawk Network enables instant gift card delivery, expanding service offerings.
Acceptance of SNAP EBT nationwide positions Gopuff as a leader in accessible grocery options.

Upsides

Collaboration with Liquid Death boosts brand visibility among younger demographics.
Aldi price match initiative attracts cost-conscious consumers, enhancing competitive edge.
'Give With Gopuff' initiative improves brand image through corporate social responsibility.

Land your dream remote job 3x faster with AI