with
t1 as(
    select
        id,
        job,
        score,
        avg(score)over(partition by job) as avg_score
    from
        grade
)

select
    id,
    job,
    score
from
    t1
where
    score>avg_score
order by
    id