Tech-tag co-occurrence

Every AI role gets tagged from a 35-term controlled vocabulary spanning agent / eval / training / inference / modality clusters. Tags that frequently appear together in the same JD pull each other close; thicker edges = more co-occurrences. The clusters that emerge organically are the real sub-disciplines of applied AI engineering right now.

All sectors · 4742Enterprise · 1275AI Frontier · 868Industrial · 650Data AI · 526Consumer · 337Banking · 239Fintech · 220Robotics · 127Defense · 115Pharma · 101Media · 73Retail · 62Hospitality · 62Telecom · 27Seattle · 27Aerospace · 14Insurance · 12Healthtech · 7

Showing 62 tagged AI roles in Retail. Layout is a Fruchterman-Reingold force simulation, run server-side to convergence.

agent_orchestration ↔ tool_use (8 JDs)agent_orchestration ↔ rag (17 JDs)agent_orchestration ↔ vector_db (8 JDs)agent_orchestration ↔ fine_tuning (7 JDs)agent_orchestration ↔ llm_observability (20 JDs)agent_orchestration ↔ guardrails (10 JDs)agent_orchestration ↔ agent_research (3 JDs)rag ↔ tool_use (8 JDs)tool_use ↔ vector_db (4 JDs)fine_tuning ↔ tool_use (4 JDs)llm_observability ↔ tool_use (6 JDs)guardrails ↔ tool_use (5 JDs)agent_research ↔ tool_use (3 JDs)rag ↔ vector_db (9 JDs)fine_tuning ↔ rag (8 JDs)llm_observability ↔ rag (17 JDs)guardrails ↔ rag (9 JDs)agent_research ↔ rag (2 JDs)fine_tuning ↔ vector_db (4 JDs)llm_observability ↔ vector_db (7 JDs)guardrails ↔ vector_db (6 JDs)agent_research ↔ vector_db (1 JDs)fine_tuning ↔ llm_observability (9 JDs)fine_tuning ↔ guardrails (5 JDs)agent_research ↔ fine_tuning (2 JDs)guardrails ↔ llm_observability (9 JDs)agent_research ↔ llm_observability (2 JDs)agent_research ↔ guardrails (1 JDs)model_serving ↔ vision (1 JDs)inference_infra ↔ vision (1 JDs)inference_infra ↔ model_serving (18 JDs)recommender_systems ↔ search_ranking (8 JDs)model_serving ↔ recommender_systems (7 JDs)inference_infra ↔ recommender_systems (4 JDs)model_serving ↔ search_ranking (9 JDs)inference_infra ↔ search_ranking (5 JDs)agent_orchestration ↔ inference_infra (8 JDs)agent_orchestration ↔ model_serving (18 JDs)inference_infra ↔ llm_observability (5 JDs)inference_infra ↔ rag (5 JDs)llm_observability ↔ model_serving (15 JDs)model_serving ↔ rag (13 JDs)fine_tuning ↔ recommender_systems (2 JDs)agent_orchestration ↔ search_ranking (2 JDs)agent_orchestration ↔ recommender_systems (2 JDs)agent_orchestration ↔ evals (10 JDs)rag ↔ search_ranking (3 JDs)llm_observability ↔ search_ranking (2 JDs)evals ↔ search_ranking (2 JDs)rag ↔ recommender_systems (2 JDs)llm_observability ↔ recommender_systems (2 JDs)evals ↔ recommender_systems (2 JDs)evals ↔ rag (7 JDs)evals ↔ llm_observability (8 JDs)evals ↔ model_serving (8 JDs)fine_tuning ↔ model_serving (6 JDs)evals ↔ fine_tuning (2 JDs)guardrails ↔ model_serving (8 JDs)evals ↔ guardrails (6 JDs)evals ↔ vector_db (3 JDs)model_serving ↔ vector_db (5 JDs)guardrails ↔ inference_infra (3 JDs)inference_infra ↔ vector_db (3 JDs)fine_tuning ↔ inference_infra (2 JDs)evals ↔ tool_use (3 JDs)model_serving ↔ tool_use (5 JDs)fine_tuning ↔ multimodal (2 JDs)llm_observability ↔ multimodal (2 JDs)multimodal ↔ rag (2 JDs)multimodal ↔ vector_db (1 JDs)inference_infra ↔ tool_use (3 JDs)evals ↔ inference_infra (3 JDs)agent_orchestration ↔ synthetic_data (1 JDs)agent_research ↔ inference_infra (1 JDs)agent_research ↔ model_serving (1 JDs)agent_research ↔ synthetic_data (1 JDs)synthetic_data ↔ tool_use (1 JDs)llm_observability ↔ synthetic_data (1 JDs)rag ↔ synthetic_data (1 JDs)fine_tuning ↔ synthetic_data (1 JDs)inference_infra ↔ synthetic_data (1 JDs)model_serving ↔ synthetic_data (1 JDs)search_ranking ↔ vector_db (1 JDs)evals ↔ multimodal (1 JDs)guardrails ↔ recommender_systems (1 JDs)guardrails ↔ search_ranking (1 JDs)guardrails ↔ multimodal (1 JDs)fine_tuning ↔ search_ranking (1 JDs)model_serving ↔ multimodal (1 JDs)multimodal ↔ recommender_systems (1 JDs)multimodal ↔ search_ranking (1 JDs)agent_orchestration ↔ code_gen (1 JDs)code_gen ↔ llm_observability (1 JDs)code_gen ↔ evals (1 JDs)code_gen ↔ guardrails (1 JDs)Agent orchestration N=34 JDs Top co-occur: LLM observability ×20 · Model serving ×18 · RAG ×17Agent orchestrationTool use N=9 JDs Top co-occur: Agent orchestration ×8 · RAG ×8 · LLM observability ×6Tool useRAG N=20 JDs Top co-occur: Agent orchestration ×17 · LLM observability ×17 · Model serving ×13RAGVector DB N=10 JDs Top co-occur: RAG ×9 · Agent orchestration ×8 · LLM observability ×7Vector DBFine-tuning N=10 JDs Top co-occur: LLM observability ×9 · RAG ×8 · Agent orchestration ×7Fine-tuningLLM observability N=25 JDs Top co-occur: Agent orchestration ×20 · RAG ×17 · Model serving ×15LLM observabilityGuardrails N=11 JDs Top co-occur: Agent orchestration ×10 · RAG ×9 · LLM observability ×9GuardrailsAgent research N=3 JDs Top co-occur: Agent orchestration ×3 · Tool use ×3 · RAG ×2Vision N=1 JDs Top co-occur: Model serving ×1 · Inference infra ×1Model serving N=31 JDs Top co-occur: Inference infra ×18 · Agent orchestration ×18 · LLM observability ×15Model servingInference infra N=18 JDs Top co-occur: Model serving ×18 · Agent orchestration ×8 · Search & ranking ×5Inference infraRecommender systems N=14 JDs Top co-occur: Search & ranking ×8 · Model serving ×7 · Inference infra ×4Recommender systemsSearch & ranking N=10 JDs Top co-occur: Model serving ×9 · Recommender systems ×8 · Inference infra ×5Search & rankingForecasting N=4 JDsEvals N=11 JDs Top co-occur: Agent orchestration ×10 · LLM observability ×8 · Model serving ×8EvalsMultimodal N=4 JDs Top co-occur: Fine-tuning ×2 · LLM observability ×2 · RAG ×2Synthetic data N=1 JDs Top co-occur: Agent orchestration ×1 · Agent research ×1 · Tool use ×1Code gen N=1 JDs Top co-occur: Agent orchestration ×1 · LLM observability ×1 · Evals ×1
18 tags · 95 co-occurrence edges · min edge weight 1. Bubble area ∝ JDs containing tag · edge thickness ∝ co-occurrence count. Hover any node for top-3 partners; click to see the JDs.

Tag velocity · last 4 weeks vs prior 4

Which technologies are hot, which are cooling. Sparkline = 12 weeks of unique roles tagged with each term, last bar on the right is this week. Sorted by absolute pickup. Tags with under 10 lifetime mentions are hidden as noise.

Tag12-week trendLast 4wPrior 4wΔTotal
agent_orchestration
0 roles
Feb 23
1 role
Mar 2
1 role
9
0 roles
16
2 roles
23
0 roles
30
1 role
Apr 6
1 role
13
5 roles
20
2 roles
27
21 roles
May 4
0 roles
11
293↑+2634
model_serving
0 roles
Feb 23
1 role
Mar 2
1 role
9
0 roles
16
1 role
23
0 roles
30
1 role
Apr 6
1 role
13
5 roles
20
1 role
27
19 roles
May 4
0 roles
11
262↑+2431
llm_observability
0 roles
Feb 23
1 role
Mar 2
1 role
9
0 roles
16
1 role
23
0 roles
30
1 role
Apr 6
1 role
13
7 roles
20
2 roles
27
10 roles
May 4
0 roles
11
202↑+1825
rag
0 roles
Feb 23
1 role
Mar 2
1 role
9
0 roles
16
0 roles
23
0 roles
30
1 role
Apr 6
1 role
13
6 roles
20
2 roles
27
7 roles
May 4
0 roles
11
161↑+1520
inference_infra
0 roles
Feb 23
0 roles
Mar 2
0 roles
9
0 roles
16
0 roles
23
0 roles
30
1 role
Apr 6
1 role
13
2 roles
20
0 roles
27
13 roles
May 4
0 roles
11
161↑+1518
recommender_systems
0 roles
Feb 23
0 roles
Mar 2
0 roles
9
0 roles
16
0 roles
23
0 roles
30
0 roles
Apr 6
2 roles
13
0 roles
20
1 role
27
9 roles
May 4
0 roles
11
120↑+1214
evals
0 roles
Feb 23
1 role
Mar 2
0 roles
9
0 roles
16
0 roles
23
0 roles
30
0 roles
Apr 6
0 roles
13
2 roles
20
1 role
27
7 roles
May 4
0 roles
11
100↑+1011
vector_db
0 roles
Feb 23
0 roles
Mar 2
0 roles
9
0 roles
16
0 roles
23
0 roles
30
0 roles
Apr 6
1 role
13
2 roles
20
0 roles
27
6 roles
May 4
0 roles
11
90↑+910
search_ranking
0 roles
Feb 23
0 roles
Mar 2
0 roles
9
0 roles
16
0 roles
23
0 roles
30
0 roles
Apr 6
0 roles
13
0 roles
20
1 role
27
7 roles
May 4
0 roles
11
80↑+810
guardrails
0 roles
Feb 23
1 role
Mar 2
1 role
9
0 roles
16
1 role
23
0 roles
30
0 roles
Apr 6
1 role
13
1 role
20
0 roles
27
6 roles
May 4
0 roles
11
81↑+711
fine_tuning
0 roles
Feb 23
0 roles
Mar 2
0 roles
9
0 roles
16
1 role
23
0 roles
30
0 roles
Apr 6
2 roles
13
3 roles
20
0 roles
27
3 roles
May 4
0 roles
11
81↑+710