Applied AI Engineer

About Bitsol

Bitsol is a product-led, AI-accelerated software studio focused on healthcare. We partner with healthcare founders and organizations in the US to design and deliver HIPAA-compliant applications through rapid, 8-week prototyping and delivery cycles. Combining deep domain experience in clinical workflows, interoperability, and compliance with modern low-code and AI-enabled development practices, we turn validated customer problems into production-ready solutions quickly and securely. Join a mission-driven, cross-functional team that prioritizes real-world impact for patients and providers.

About the RoleAs an Applied AI Engineer you will design, build, evaluate, and operate agentic AI systems and retrieval‑augmented pipelines for HIPAA‑sensitive healthcare products. You will own production integration of LLMs and multimodal models, the orchestration of multi‑agent workflows, and the observability, safety and privacy guarantees required for clinical and regulated environments. You will work cross‑functionally with Product, Engineering, DevOps and Compliance to ship reliable, auditable AI features.

Key Responsibilities

• Design and implement agentic systems: multi‑step agents, tool/connector interfaces, action selection policies, orchestrators and workflow engines.
• Build and maintain RAG pipelines: document ingestion, chunking strategies, embeddings, vector indexes (FAISS/Pinecone/Weaviate), semantic search and recall tuning.
• Integrate LLMs and multimodal models (OpenAI/Anthropic/local models/LLM-as-a-service) with robust prompt templates, function‑calling/tool interfaces and orchestration frameworks.
• Implement safety, alignment and privacy controls: PII/PHI scrubbing, role/row access patterns, policy enforcement, content moderation and red‑teaming.
• Produce reproducible evaluation and observability: automated test suites, evaluation metrics (faithfulness, precision, ROUGE/EM where applicable), logging, drift detection and dashboards.
• Harden production services: API design, auth, batching, caching, rate limiting, containerization (Docker), CI/CD, SLOs, scaling and secure hosting patterns suitable for HIPAA workloads.
• Build connectors and integrations (EHR APIs, FHIR/HL7, internal knowledge bases), define access & consent flows, and ensure audit trails for PHI access.
• Define and maintain prompt/agent libraries, tooling contracts (input/output schemas), and documentation for reproducibility and handoff.
• Collaborate with Product and Compliance to translate product requirements into agent behaviors and acceptance criteria; participate in incident response and post‑mortems.

Key Qualifications

• 4+ years in applied ML/AI engineering or software engineering with significant AI responsibilities; proven production experience.
• Demonstrable experience building agentic systems (tool‑using agents, multi‑step orchestrations, or similar), not limited to single‑shot prompting.
• Strong, hands‑on experience with retrieval‑augmented systems: embeddings, vector DBs (FAISS/Pinecone/Weaviate), chunking/indexing strategies and retrieval tuning.
• Proven integration experience with LLMs and model APIs, including prompt engineering, function calling, and orchestration (e.g., LangChain, LangGraph, CrewAI or equivalent).
• Production engineering skills: Python, async services, API design, Docker, CI/CD pipelines, and familiarity with deployment patterns for low‑latency/high‑availability services.
• Practical knowledge of evaluation & observability for LLM systems: logging request/response, evaluation pipelines, metrics, and drift detection.
• Strong understanding of data privacy, security and PHI/PII handling; experience designing for auditability and HIPAA compliance requirements.
• Experience debugging model/system failures and performing root‑cause analysis for hallucinations, retrieval failures, and agent misbehavior.
• Solid software engineering fundamentals: testing, reproducibility, documentation and collaboration in cross‑functional teams.
• Clear communicator with product sensibility and ability to translate product needs into technical designs.

Preferred / Ideal Candidate

• Experience with healthcare data, EHR integrations and interoperability standards (FHIR, HL7).
• Hands‑on experience deploying or fine‑tuning LLMs (LoRA, SFT, full model deployments) and working with on‑prem/local model deployments.
• Familiarity with MLOps and feature/serving platforms (MLflow, BentoML, Tecton, Feast, feature stores).
• Experience with agent orchestration/workflow tools (Temporal, Airflow, CrewAI, LangGraph) and multi‑agent coordination patterns.
• Prior work on model safety, alignment tooling, content moderation pipelines and red‑teaming exercises.
• Experience with knowledge graphs, schema/ontology design, or hybrid retrieval (BM25 + embeddings + re‑ranking).
• Background in building evaluation frameworks and operationalizing model QA in regulated environments.
• Mentorship experience and a desire to build reusable prompt/agent libraries and team best practices.

Perks and Benefits:

• Biannual salary increment
• Health and fitness allowance
• Company-sponsored sports and adventure activities
• Bonuses
• Fuel Allowance
• Awesome learning environment
• Growth opportunities

Job Type: Full-time

Ability to commute/relocate:

• Islamabad: Reliably commute or planning to relocate before starting work (Required)

Experience:

• applied ML/AI engineering : 4 years (Required)

Work Location: In person