Job Description

Roles & Responsibilities:

Architect scalable AI solutions:

Define end-to-end reference architectures (LLM/RAG, NLP, vision, agentic workflows) that move cleanly from

DemoBytes ? POC ? MVP ? Demoable ? Production

.

Own full-stack delivery:

Build features across data/ML, backend APIs/services (FastAPI/Flask), and lightweight UIs (React/Next.js) for demoable, user-ready outputs.

Rapid prototyping:

Stand up POCs in days; harden validated solutions into MVP and production with incremental quality/security gates.

MLOps & platformization:

Implement CI/CD/CT for models, datasets, prompts; automate evals, canary/rollback, versioning, model/data drift monitoring, and experiment tracking (W&B/MLflow).

Integration & interoperability:

Embed AI into existing products and workflows via APIs, queues, SDKs, and webhooks with clear SLAs and observability.

Operate what you build:

Instrument services, track p95 latency/availability/cost, and drive continuous improvement post-launch.

Mentor & uplift:

Coach engineers on best practices (prompting, vector design, evals, latency/cost tuning, secure data handling).

Release cadence:

Maintain

monthly demo releases

and

production releases every two months

with ALM-driven governance.

Ethical AI & compliance:

Apply privacy-by-design, bias testing/mitigation, model cards, auditability, and data protection controls; ensure documentation in ALM.

Trendwatching:

Track state-of-the-art AI (models, toolchains, infra) and pragmatically incorporate breakthroughs into roadmaps.

Qualifications:

4-6 years

delivering AI/ML features

to production

with fast

POC ? MVP ? Production

cycles. Strong ML/DL fundamentals; hands-on with

PyTorch

and/or

TensorFlow/Keras

; LLMs (prompting, fine-tuning/LoRA), RAG patterns, and evaluation.

Python

proficiency; scikit-learn, spaCy/NLTK;

Hugging Face

(Transformers/Datasets/PEFT); familiarity with

YOLO

/FastAI (role-relevant). Backend engineering for production (FastAPI/Flask), auth, caching, testing; practical

React/Next.js

for demoable UIs.

MLOps:

Docker/Kubernetes, CI/CD (GitHub Actions/Azure DevOps/Jenkins), experiment tracking (

Weights & Biases

/MLflow), monitoring (Prometheus/Grafana/OpenTelemetry). Data & storage: SQL/NoSQL (Postgres, Redis), object stores; vector DBs (FAISS/Milvus/pgvector) and retrieval design. Cloud: AWS/Azure/GCP with cost/latency/performance trade-off literacy.

AI productivity tools (required):

Cursor, Windsurf, Claude, Copilot

for accelerated prototyping, code gen/review, and prompt workflows. Effective communication; crisp documentation and governance in ALM. Working knowledge of ethical AI and

data protection

(PII handling, access controls, audit trails).
Job Type: Full-time

Experience:

AI: 4 years (Required) Python: 4 years (Required)
Work Location: In person