AI Developer ( Uzbekistan, Kazakhstan, Kyrgystan)

Компания "EVERTECH"

About Us

Evertech helps AI & SaaS startups scale quickly without burning capital. We build lean, high-impact engineering pods that ship Gen-AI features used by millions. You’ll join a tight-knit crew that moves fast, owns the outcome, and treats LLMs as first-class building blocks rather than magic black boxes.

About the Role

You will design, build, and productize features that harness large-language models (LLMs) and multimodal models for a leading financial services company in New York. Language choice isn’t a blocker—our stack today is mostly Python/TypeScript, but we care far more about your ability to reason about model capabilities, data flows, and user value.

Responsibilities

• Prototype and ship RAG, agent, or generative-UX features using frameworks like LangChain or LlamaIndex.
• Select and integrate model endpoints (OpenAI GPT-4o, Anthropic Claude, Gemini 1.5, custom fine-tunes on Bedrock, etc.).
• Design retrieval layers—embedding generation, chunking strategies, and vector-DB schemas (Pinecone, Weaviate, Chroma, Milvus).
• Implement evaluation & monitoring pipelines with LangSmith, RAGAS, or comparable LLMOps tools to guardrail quality and cost.
• Automate deployments via containerized CI/CD on AWS/GCP/Kubernetes or TrueFoundry.
• Write clean, testable code; add tracing & logging around LLM calls.
• Optimize prompt chains for latency, token-use, and factuality.
• Collaborate with design & product to craft user-facing AI experiences (chat widgets, autocomplete, code-assist, etc.).
• Conduct ablation studies, A/B tests, and data-driven iterations.

Requirements

• 3–5 years professional software development; at least 1 year hands-on with LLMs or generative models.
• Proven delivery of at least one production Gen-AI feature (chatbot, summarizer, code-assist, content-gen, etc.).
• Deep familiarity with LangChain or LlamaIndex and at least one vector DB (Pinecone / Weaviate / Chroma / Milvus).
• Comfort orchestrating models on AWS Bedrock, Azure OpenAI, Vertex AI, or Hugging Face APIs.
• Solid skills in Python or TypeScript plus tooling (FastAPI/Express, pytest/jest, Docker).
• Experience securing and scaling cloud workloads (IaC, OAuth, VPCs).
• Track record of data-driven experimentation & prompt engineering.
• Excellent written & verbal English; able to explain complex AI behavior to non-experts.

Nice-to-Have

• Fine-tuning/LoRA/QLoRA on open-weight models.
• Knowledge of multimodal models (image, audio).
• Familiarity with LLM security red-teaming and policy enforcement.
• Prior startup or dev-tool product experience.

What We Offer

• Competitive USD-pegged compensation + stock/bonus.
• Fully remote (or Tashkent HQ) with flexible hours.
• Annual learning budget for conferences, compute credits, and courses.
• Top-of-line hardware & generous open-source contributions policy.
• A no-bureaucracy culture where your Gen-AI ideas hit production in weeks, not quarters.