AI / LLM Engineer + Deployment Engineer
About the Role
We are hiring an AI / LLM Engineer to own AI-powered features for the Digital Marketing function end-to-end — content automation, AI-driven analytics, chatbots on our websites, and LLM integrations into our admin panel — and to independently handle all deployments for DM-owned tools and websites. The candidate will be the DM department's dedicated AI and deployment engineer, reporting fully to the DM Leader, and will not depend on any other team for routine releases, server access, or production fixes for DM properties.
Key Responsibilities
AI & LLM (60%)
- Build and own LLM-powered features for Harshwal's websites and DM admin panel — chatbots, content generators, SEO drafting agents, summarizers, hashtag and caption generators, post-performance analyzers
- Integrate APIs from OpenAI, Anthropic Claude, and AWS Bedrock into Python applications
- Build prompt engineering libraries, prompt versioning, and prompt evaluation
- Implement RAG (Retrieval Augmented Generation) pipelines — embeddings, vector databases (pgvector, FAISS, Chroma)
- Track token usage and cost per project; submit weekly cost reports to DM Leader
- Integrate LLM-based insights into social media analytics dashboards
- Automate content QA — brand voice, grammar, and tone checks before publishing
Deployment & DevOps (40%)
- Deploy DM-owned tools and websites on Linux servers — Ubuntu Digital Ocean instances (BLR1, SFO3, NYC zones) and assigned internal servers
- Configure SSH access, port mapping (5000–9000 range for internal services), systemd services, Nginx reverse proxy
- Manage Docker containers where applicable
- Maintain deployment documentation and runbooks for every DM-owned project deployed
- Independently handle production fixes, rollbacks, and uptime monitoring for all DM properties
- Manage SSL certificates, domain and sub-domain routing, and DNS for DM-owned web properties
Mandatory Technical Skills
- Python: FastAPI or Flask — 1+ year hands-on
- LLM integration: Direct API experience with at least one of — OpenAI, Anthropic Claude, AWS Bedrock, Azure OpenAI
- Prompt engineering: Structured prompts, system messages, function / tool calling
- Linux deployment: Ubuntu, SSH, port management, systemd, Nginx, Let's Encrypt SSL
- Server environments: Digital Ocean droplets (1 – 32 GB RAM, BLR / NYC / SFO zones), AWS EC2, and internal Linux servers
